fix: force offline mode when loading cached models (Qwen TTS & Whisper)

Qwen TTS and Whisper Base make network calls to HuggingFace even when
model weights are fully cached locally, because from_pretrained()
defaults to local_files_only=False. This causes failures for offline
users.

Add a reusable force_offline_if_cached() context manager that sets
HF_HUB_OFFLINE=1 during model loading when is_model_cached() is True.
Applied to all four affected load paths:

- PyTorchTTSBackend (Qwen TTS)
- PyTorchSTTBackend (Whisper)
- MLXTTSBackend (refactored from inline implementation)
- MLXSTTBackend (previously unprotected)

Closes #82

.agents/skills/add-tts-engine/SKILL.md

@@ -0,0 +1,120 @@
+---
+name: add-tts-engine
+description: Use this skill to add a new TTS engine to Voicebox. It walks through dependency research, backend implementation, frontend wiring, PyInstaller bundling, and frozen-build testing. Always start with Phase 0 (dependency audit) before writing any code.
+---
+
+# Add TTS Engine
+
+## Goal
+
+Integrate a new text-to-speech engine into Voicebox end-to-end: dependency research, backend protocol implementation, frontend UI wiring, PyInstaller bundling, and frozen-build verification. The user should only need to test the final build locally.
+
+## Reference Doc
+
+The full phased guide lives at `docs/content/docs/developer/tts-engines.mdx`. **Read this file in its entirety before starting.** It contains:
+
+- Phase 0: Dependency research (mandatory before writing code)
+- Phase 1: Backend implementation (`TTSBackend` protocol)
+- Phase 2: Route and service integration (usually zero changes)
+- Phase 3: Frontend integration (5 files)
+- Phase 4: Dependencies (`requirements.txt`, justfile, CI, Docker)
+- Phase 5: PyInstaller bundling (`build_binary.py` + `server.py`)
+- Phase 6: Common upstream workarounds
+- Implementation checklist (gate between phases)
+
+## Workflow
+
+### 1. Read the guide
+
+```bash
+# Read the full TTS engines doc
+cat docs/content/docs/developer/tts-engines.mdx
+```
+
+Internalize all phases, especially Phase 0 and Phase 5. The v0.2.3 release was three patch releases because Phase 0 was skipped.
+
+### 2. Dependency research (Phase 0)
+
+Clone the model library into a temporary directory and audit it. Do NOT skip this.
+
+```bash
+mkdir /tmp/engine-research && cd /tmp/engine-research
+git clone <model-library-url>
+```
+
+Run the grep searches from Phase 0.2 in the guide against the cloned source and its transitive dependencies. Produce a written dependency audit covering:
+
+1. PyPI vs non-PyPI packages
+2. PyInstaller directives needed (`--collect-all`, `--copy-metadata`, `--hidden-import`)
+3. Runtime data files that must be bundled
+4. Native library paths that need env var overrides in frozen builds
+5. Monkey-patches needed (`torch.load`, float64, MPS, HF token)
+6. Sample rate
+7. Model download method (`from_pretrained` vs `snapshot_download` + `from_local`)
+
+Test model loading and generation on CPU in the throwaway venv before proceeding.
+
+### 3. Implement (Phases 1–4)
+
+Follow the guide's phases in order. Key files to modify:
+
+**Backend (Phase 1):**
+- Create `backend/backends/<engine>_backend.py`
+- Register in `backend/backends/__init__.py` (ModelConfig + TTS_ENGINES + factory)
+- Update regex in `backend/models.py`
+
+**Frontend (Phase 3):**
+- `app/src/lib/api/types.ts` — engine union type
+- `app/src/lib/constants/languages.ts` — ENGINE_LANGUAGES
+- `app/src/components/Generation/EngineModelSelector.tsx` — ENGINE_OPTIONS, ENGINE_DESCRIPTIONS
+- `app/src/lib/hooks/useGenerationForm.ts` — Zod schema, model-name mapping
+- `app/src/components/ServerSettings/ModelManagement.tsx` — MODEL_DESCRIPTIONS
+
+**Dependencies (Phase 4):**
+- `backend/requirements.txt`
+- `justfile` (setup-python, setup-python-release targets)
+- `.github/workflows/release.yml`
+- `Dockerfile` (if applicable)
+
+### 4. PyInstaller bundling (Phase 5)
+
+Register the engine in `backend/build_binary.py`:
+- `--hidden-import` for the backend module and model package
+- `--collect-all` for packages using `inspect.getsource`, shipping data files, or native libraries
+- `--copy-metadata` for packages using `importlib.metadata`
+
+If the engine has native data paths, add `os.environ.setdefault()` in `backend/server.py` inside the `if getattr(sys, 'frozen', False):` block.
+
+### 5. Verify in dev mode
+
+```bash
+just dev
+```
+
+Test the full chain: model download → load → generate → voice cloning.
+
+### 6. Use the checklist
+
+Walk through the Implementation Checklist at the bottom of `tts-engines.mdx`. Every item must be checked before handing the build to the user.
+
+## Key Lessons (from v0.2.3)
+
+These are the most common failure modes. Phase 0 research catches all of them:
+
+| Pattern | Symptom in Frozen Build | Fix |
+|---------|------------------------|-----|
+| `@typechecked` / `inspect.getsource()` | "could not get source code" | `--collect-all <package>` |
+| Package ships pretrained model files | `FileNotFoundError` for `.pth.tar`, `.yaml` | `--collect-all <package>` |
+| C library with hardcoded system paths | `FileNotFoundError` for `/usr/share/...` | `--collect-all` + env var in `server.py` |
+| `importlib.metadata.version()` | "No package metadata found" | `--copy-metadata <package>` |
+| `torch.load` without `map_location` | CUDA device not available on CPU build | Monkey-patch `torch.load` |
+| `torch.from_numpy` on float64 data | dtype mismatch RuntimeError | Cast to `.float()` |
+| `token=True` in HF download calls | Auth failure without stored HF token | Use `snapshot_download(token=None)` + `from_local()` |
+
+## Notes
+
+- The route and service layers have zero per-engine dispatch points. `main.py` requires zero changes.
+- The model config registry in `backends/__init__.py` handles all dispatch automatically.
+- Use `get_torch_device()` and `model_load_progress()` from `backends/base.py` — don't reimplement device detection or progress tracking.
+- Always test with a **clean HuggingFace cache** (no pre-downloaded models from dev).
+- Do NOT push or create a release. Hand the build to the user for local testing.

.agents/skills/draft-release-notes/SKILL.md

@@ -0,0 +1,94 @@
+---
+name: draft-release-notes
+description: Use this skill to draft or update the [Unreleased] section of CHANGELOG.md from the actual changes since the last tag. Run this at any point during development to keep a working copy of the release narrative. Does NOT bump versions or create tags.
+---
+
+# Draft Release Notes
+
+## Goal
+
+Update the `[Unreleased]` section at the top of `CHANGELOG.md` with a narrative release story based on the real changes since the last tag. This is a **non-destructive working copy** — run it as many times as you want during development.
+
+## Workflow
+
+1. **Identify the last release tag and gather changes.**
+
+   ```bash
+   LAST_TAG=$(git tag --list "v*" --sort=-v:refname | head -n 1)
+   echo "Last tag: $LAST_TAG"
+   ```
+
+   Then collect raw material from three sources:
+
+   a. **Commit log since last tag:**
+   ```bash
+   git log --oneline "$LAST_TAG"..HEAD
+   ```
+
+   b. **GitHub-generated release notes preview** (PR titles, new contributors):
+   ```bash
+   gh api repos/:owner/:repo/releases/generate-notes \
+     -f tag_name="vNEXT" \
+     -f target_commitish="$(git rev-parse HEAD)" \
+     -f previous_tag_name="$LAST_TAG" \
+     --jq '.body'
+   ```
+
+   c. **Diff stat for theme analysis:**
+   ```bash
+   git diff --stat "$LAST_TAG"..HEAD
+   ```
+
+2. **Draft the release narrative.**
+
+   Write markdown for the `[Unreleased]` section following the format below. Do not include the `## [Unreleased]` heading itself — just the body content.
+
+3. **Update CHANGELOG.md.**
+
+   Replace everything between `## [Unreleased]` and the next `## [` heading with the new draft. Preserve the HTML comment header and all existing release sections below.
+
+   The `[Unreleased]` section must always exist and always be the first section after the header comments.
+
+4. **Do NOT commit, tag, or bump versions.** Just leave the file modified in the working tree.
+
+## Release Story Format
+
+Structure the `[Unreleased]` section like this:
+
+```markdown
+## [Unreleased]
+
+<One strong opening paragraph: what this release is about and why it matters.
+Tie it to concrete shipped changes. No vague hype.>
+
+<One paragraph on major technical shifts, if applicable.>
+
+### <Feature/Theme Group>
+- Bullet points with specifics
+- Reference PRs where available: ([#123](https://github.com/jamiepine/voicebox/pull/123))
+
+### <Another Group>
+- ...
+
+### Bug Fixes
+- ...
+```
+
+### Style Guidelines
+
+- **Factual and specific.** Every claim should trace to a real commit or PR.
+- **Narrative over list.** Lead with paragraphs that tell the story, then support with bullets.
+- **Group by theme, not by commit.** Cluster related changes under descriptive headings.
+- **Reference PRs** where they exist, but don't fabricate them.
+- **Skip trivial chores** (typo fixes, CI tweaks) unless they're the bulk of the release.
+- **Match the voice of existing releases** — look at the v0.2.1 and v0.2.3 entries in CHANGELOG.md for tone reference.
+
+## When There Are No Changes
+
+If `git log "$LAST_TAG"..HEAD` is empty, leave the `[Unreleased]` section empty (just the heading) and tell the user there's nothing to draft.
+
+## Notes
+
+- This skill only touches the `[Unreleased]` section. It never modifies stamped release sections.
+- The agent can be asked to run this skill at any point — mid-feature, before a PR, or right before cutting a release.
+- The `release-bump` skill depends on this draft being up to date before it finalizes.

.agents/skills/release-bump/SKILL.md

@@ -0,0 +1,124 @@
+---
+name: release-bump
+description: Use this skill to finalize a release. It stamps the [Unreleased] changelog section with a version and date, runs bumpversion to update all version files, and creates the release commit and tag. Only run this when you're ready to ship.
+---
+
+# Release Bump
+
+## Goal
+
+Finalize the changelog draft, bump the version across all tracked files, and create a tagged release commit. After this skill runs, the repo has a clean release commit and tag ready to push.
+
+## Prerequisites
+
+- `gh` CLI installed and authenticated (`gh auth status`).
+- `bumpversion` installed (`pip install bumpversion` or available in the project venv).
+- The `[Unreleased]` section of `CHANGELOG.md` should already contain the release narrative. If it's empty or stale, run the `draft-release-notes` skill first.
+
+## Workflow
+
+1. **Verify the working tree is clean** (except `CHANGELOG.md` which may have the draft).
+
+   ```bash
+   git status --porcelain
+   ```
+
+   Only `CHANGELOG.md` (and optionally `.agents/` files) should be modified. If there are other uncommitted changes, stop and ask the user to commit or stash them first.
+
+2. **Determine the bump level.**
+
+   Ask the user if not specified: `patch`, `minor`, or `major`. Check the current version:
+
+   ```bash
+   grep '^current_version' .bumpversion.cfg
+   ```
+
+3. **Stamp the changelog.**
+
+   Read the current `[Unreleased]` content from `CHANGELOG.md`. Compute the new version (based on bump level and current version). Then:
+
+   a. Replace the `## [Unreleased]` section body with an empty placeholder.
+   b. Insert a new stamped section immediately after `## [Unreleased]`:
+
+   ```markdown
+   ## [Unreleased]
+
+   ## [X.Y.Z] - YYYY-MM-DD
+
+   <the content that was in [Unreleased]>
+   ```
+
+   c. Update the reference links at the bottom of the file:
+   - Change the `[Unreleased]` link to compare against the new tag
+   - Add a new link for the new version
+
+   ```markdown
+   [Unreleased]: https://github.com/jamiepine/voicebox/compare/vX.Y.Z...HEAD
+   [X.Y.Z]: https://github.com/jamiepine/voicebox/compare/vPREVIOUS...vX.Y.Z
+   ```
+
+4. **Stage the changelog.**
+
+   ```bash
+   git add CHANGELOG.md
+   ```
+
+5. **Run bumpversion.**
+
+   ```bash
+   bumpversion --allow-dirty <patch|minor|major>
+   ```
+
+   The `--allow-dirty` flag is needed because `CHANGELOG.md` is already staged. bumpversion will:
+   - Update version strings in all tracked files (see `.bumpversion.cfg`)
+   - Create a commit with message `Bump version: X.Y.Z -> A.B.C`
+   - Create a tag `vA.B.C`
+
+   The staged `CHANGELOG.md` will be included in this commit automatically.
+
+6. **Verify results.**
+
+   ```bash
+   git show --name-only --stat HEAD
+   git tag --list "v*" --sort=-v:refname | head -n 5
+   ```
+
+   Confirm the commit contains:
+   - `CHANGELOG.md`
+   - `.bumpversion.cfg`
+   - `tauri/src-tauri/tauri.conf.json`
+   - `tauri/src-tauri/Cargo.toml`
+   - `package.json`
+   - `app/package.json`
+   - `tauri/package.json`
+   - `landing/package.json`
+   - `web/package.json`
+   - `backend/__init__.py`
+
+   Confirm the new tag exists.
+
+7. **Do NOT push** unless the user explicitly asks. Report the tag name and suggest:
+
+   ```
+   Ready to push. When you're ready:
+     git push origin main --follow-tags
+   ```
+
+## Version Calculation Reference
+
+Given current version `X.Y.Z`:
+- `patch` -> `X.Y.(Z+1)`
+- `minor` -> `X.(Y+1).0`
+- `major` -> `(X+1).0.0`
+
+## Error Recovery
+
+- If bumpversion fails, the tag won't exist. Fix the issue and re-run — bumpversion is idempotent as long as the tag doesn't already exist.
+- If you need to undo a release commit (before pushing): `git tag -d vX.Y.Z && git reset --soft HEAD~1`
+- Never amend a release commit that has been pushed.
+
+## Notes
+
+- When the tag is pushed, the release CI (`.github/workflows/release.yml`) automatically extracts the matching version section from `CHANGELOG.md` and uses it as the GitHub Release body. No manual copy-paste needed.
+- The release commit message is controlled by `.bumpversion.cfg` (`Bump version: X.Y.Z -> A.B.C`). Do not override it.
+- If you need to manually update the GitHub Release body after the fact: `gh release edit vX.Y.Z --notes-file <(sed -n '/## \[X.Y.Z\]/,/## \[/p' CHANGELOG.md | head -n -1)`

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.2.3
+current_version = 0.3.1
 commit = True
 tag = True
 tag_name = v{new_version}

.github/workflows/release.yml

@@ -62,6 +62,7 @@ jobs:
           pip install pyinstaller
           pip install -r backend/requirements.txt
           pip install --no-deps chatterbox-tts
+          pip install --no-deps hume-tada
 
       - name: Install MLX dependencies (Apple Silicon only)
         if: matrix.backend == 'mlx'
@@ -123,6 +124,29 @@ jobs:
           p12-file-base64: ${{ secrets.APPLE_CERTIFICATE }}
           p12-password: ${{ secrets.APPLE_CERTIFICATE_PASSWORD }}
 
+      - name: Extract release notes from CHANGELOG.md
+        id: changelog
+        shell: bash
+        run: |
+          # Get the version from the tag (strip leading 'v')
+          VERSION="${GITHUB_REF_NAME#v}"
+
+          # Extract the section for this version from CHANGELOG.md
+          # Matches from "## [X.Y.Z]" until the next "## [" heading
+          NOTES=$(sed -n "/^## \[${VERSION}\]/,/^## \[/{/^## \[${VERSION}\]/d;/^## \[/d;p;}" CHANGELOG.md)
+
+          # Fall back to a placeholder if the version isn't in the changelog
+          if [ -z "$(echo "$NOTES" | tr -d '[:space:]')" ]; then
+            NOTES="See the assets below to download and install this version."
+          fi
+
+          # Use multiline output syntax
+          {
+            echo "notes<<CHANGELOG_EOF"
+            echo "$NOTES"
+            echo "CHANGELOG_EOF"
+          } >> "$GITHUB_OUTPUT"
+
       - uses: tauri-apps/[email protected]
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
@@ -139,17 +163,7 @@ jobs:
           projectPath: tauri
           tagName: v__VERSION__
           releaseName: "voicebox v__VERSION__"
-          releaseBody: |
+          releaseBody: ${{ steps.changelog.outputs.notes }}
-            ## What's Changed
-            See the assets below to download and install this version.
-
-            ### Installation
-            - **macOS (Apple Silicon)**: Download the `aarch64.dmg` file - uses MLX for fast native inference
-            - **macOS (Intel)**: Download the `x64.dmg` file - uses PyTorch
-            - **Windows**: Download the `.msi` installer
-            - **Linux**: Compile from source (see README)
-
-            The app includes automatic updates - future updates will be installed automatically.
           releaseDraft: true
           prerelease: false
           args: ${{ matrix.args }}
@@ -175,43 +189,48 @@ jobs:
           pip install pyinstaller
           pip install -r backend/requirements.txt
           pip install --no-deps chatterbox-tts
+          pip install --no-deps hume-tada
 
-      - name: Install PyTorch with CUDA 12.1
+      - name: Install PyTorch with CUDA 12.8
         run: |
-          pip install torch --index-url https://download.pytorch.org/whl/cu121 --force-reinstall --no-deps
+          pip install torch --index-url https://download.pytorch.org/whl/cu128 --force-reinstall --no-deps
-          pip install torchaudio --index-url https://download.pytorch.org/whl/cu121
+          pip install torchaudio --index-url https://download.pytorch.org/whl/cu128 --force-reinstall --no-deps
 
       - name: Verify CUDA support in torch
         run: |
           python -c "import torch; print(f'CUDA available in build: {torch.cuda.is_available()}'); print(f'CUDA version: {torch.version.cuda}')"
 
-      - name: Build CUDA server binary
+      - name: Build CUDA server binary (onedir)
         shell: bash
         working-directory: backend
         run: python build_binary.py --cuda
 
-      - name: Split binary for GitHub Releases
+      - name: Package into server core + CUDA libs archives
         shell: bash
         run: |
-          python scripts/split_binary.py \
+          python scripts/package_cuda.py \
-            backend/dist/voicebox-server-cuda.exe \
+            backend/dist/voicebox-server-cuda/ \
-            --output release-assets/
+            --output release-assets/ \
+            --cuda-libs-version cu128-v1 \
+            --torch-compat ">=2.7.0,<2.11.0"
 
-      - name: Upload split parts to GitHub Release
+      - name: Upload archives to GitHub Release
         if: startsWith(github.ref, 'refs/tags/')
-        uses: softprops/action-gh-release@v1
+        uses: softprops/action-gh-release@v2
         with:
           files: |
-            release-assets/voicebox-server-cuda.part*.exe
+            release-assets/voicebox-server-cuda.tar.gz
-            release-assets/voicebox-server-cuda.sha256
+            release-assets/voicebox-server-cuda.tar.gz.sha256
-            release-assets/voicebox-server-cuda.manifest
+            release-assets/cuda-libs-cu128-v1.tar.gz
+            release-assets/cuda-libs-cu128-v1.tar.gz.sha256
+            release-assets/cuda-libs.json
           draft: true
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
 
-      - name: Upload binary as workflow artifact
+      - name: Upload onedir as workflow artifact
         uses: actions/upload-artifact@v4
         with:
           name: voicebox-server-cuda-windows
-          path: backend/dist/voicebox-server-cuda.exe
+          path: backend/dist/voicebox-server-cuda/
           retention-days: 7

.gitignore

@@ -50,6 +50,14 @@ logs/
 app/openapi.json
 tauri/src-tauri/binaries/*
 tauri/src-tauri/gen/Assets.car
+tauri/src-tauri/gen/voicebox.icns
+tauri/src-tauri/gen/partial.plist
+
+# PyInstaller
+*.spec
+
+# Windows artifacts
+nul
 
 # Temporary
 tmp/

CHANGELOG.md

@@ -1,96 +1,464 @@
+<!-- This file is compiled automatically during the release workflow. -->
+<!-- Do not edit manually — your changes will be overwritten. -->
+<!-- To update the draft: ask the agent to use the draft-release-notes skill. -->
+<!-- To finalize a release: ask the agent to use the release-bump skill. -->
+
 # Changelog
 
-All notable changes to Voicebox will be documented in this file.
-
-The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
-and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
-
 ## [Unreleased]
 
-### Fixed
+## [0.3.0] - 2026-03-17
-- **Profile Name Validation** - Added proper validation to prevent duplicate profile names ([#134](https://github.com/jamiepine/voicebox/issues/134))
-  - Users now receive clear error messages when attempting to create or update profiles with duplicate names
-  - Improved error handling in create and update profile API endpoints
-  - Added comprehensive test suite for duplicate name validation
 
-## [0.1.0] - 2026-01-25
+This release rewrites the backend into a modular architecture, overhauls the settings UI into routed sub-pages, fixes audio player freezing, migrates documentation to Fumadocs, and ships a batch of bug fixes targeting the most-reported issues from the tracker.
 
-### Added
+The backend's 3,000-line monolith `main.py` has been decomposed into domain routers, a services layer, and a proper database package. A style guide and ruff configuration now enforce consistency. On the frontend, settings have been split into dedicated routed pages with server logs, a changelog viewer, and an about page. The audio player no longer freezes mid-playback, and model loading status is now visible in the UI. Seven user-reported bugs have been fixed, including server crashes during sample uploads, generation list staleness, cryptic error messages, and CUDA support for RTX 50-series GPUs.
 
-#### Core Features
+### Settings Overhaul ([#294](https://github.com/jamiepine/voicebox/pull/294))
-- **Voice Cloning** - Clone voices from audio samples using Qwen3-TTS (1.7B and 0.6B models)
+- Split settings into routed sub-tabs: General, Generation, GPU, Logs, Changelog, About
-- **Voice Profile Management** - Create, edit, and organize voice profiles with multiple samples
+- Added live server log viewer with auto-scroll
-- **Speech Generation** - Generate high-quality speech from text using cloned voices
+- Added in-app changelog page that parses `CHANGELOG.md` at build time
-- **Generation History** - Track all generations with search and filtering capabilities
+- Added About page with version info, license, and generation folder quick-open
-- **Audio Transcription** - Automatic transcription powered by Whisper
+- Extracted reusable `SettingRow` component for consistent setting layouts
-- **In-App Recording** - Record audio samples directly in the app with waveform visualization
 
-#### Desktop App
+### Audio Player Fix ([#293](https://github.com/jamiepine/voicebox/pull/293))
-- **Tauri Desktop App** - Native desktop application for macOS, Windows, and Linux
+- Fixed audio player freezing during playback
-- **Local Server Mode** - Embedded Python server runs automatically
+- Improved playback UX with better state management and listener cleanup
-- **Remote Server Mode** - Connect to a remote Voicebox server on your network
+- Fixed restart race condition during regeneration
-- **Auto-Updates** - Automatic update notifications and installation
+- Added stable keys for audio element re-rendering
+- Improved accessibility across player controls
 
-#### API
+### Backend Refactor ([#285](https://github.com/jamiepine/voicebox/pull/285))
-- **REST API** - Full REST API for voice synthesis and profile management
+- Extracted all routes from `main.py` into 13 domain routers under `backend/routes/` — `main.py` dropped from ~3,100 lines to ~10
-- **OpenAPI Documentation** - Interactive API docs at `/docs` endpoint
+- Moved CRUD and service modules into `backend/services/`, platform detection into `backend/utils/`
-- **Type-Safe Client** - Auto-generated TypeScript client from OpenAPI schema
+- Split monolithic `database.py` into a `database/` package with separate `models`, `session`, `migrations`, and `seed` modules
+- Added `backend/STYLE_GUIDE.md` and `pyproject.toml` with ruff linting config
+- Removed dead code: unused `_get_cuda_dll_excludes`, stale `studio.py`, `example_usage.py`, old `Makefile`
+- Deduplicated shared logic across TTS backends into `backends/base.py`
+- Improved startup logging with version, platform, data directory, and database stats
+- Fixed startup database session leak — sessions now rollback and close in `finally` block
+- Isolated shutdown unload calls so one backend failure doesn't block the others
+- Handled null duration in `story_items` migration
+- Reject model migration when target is a subdirectory of source cache
 
-#### Technical
+### Documentation Rewrite ([#288](https://github.com/jamiepine/voicebox/pull/288))
-- **Voice Prompt Caching** - Fast regeneration with cached voice prompts
+- Migrated docs site from Mintlify to Fumadocs (Next.js-based)
-- **Multi-Sample Support** - Combine multiple audio samples for better voice quality
+- Rewrote introduction and root page with content from README
-- **GPU/CPU/MPS Support** - Automatic device detection and optimization
+- Added "Edit on GitHub" links and last-updated timestamps on all pages
-- **Model Management** - Lazy loading and VRAM management
+- Generated OpenAPI spec and auto-generated API reference pages
-- **SQLite Database** - Local data persistence
+- Removed stale planning docs (`CUDA_BACKEND_SWAP`, `EXTERNAL_PROVIDERS`, `MLX_AUDIO`, `TTS_PROVIDER_ARCHITECTURE`, etc.)
+- Sidebar groups now expand by default; root redirects to `/docs`
+- Added OG image metadata and `/og` preview page
 
-### Technical Details
+### UI & Frontend
+- Added model loading status indicator and effects preset dropdown ([3187344](https://github.com/jamiepine/voicebox/commit/3187344))
+- Fixed take-label race condition during regeneration
+- Added accessible focus styling to select component
+- Softened select focus indicator opacity
+- Addressed 4 critical and 12 major issues from CodeRabbit review
 
-- Built with Tauri v2 (Rust + React)
+### Bug Fixes ([#295](https://github.com/jamiepine/voicebox/pull/295))
-- FastAPI backend with async Python
+- Fixed sample uploads crashing the server — audio decoding now runs in a thread pool instead of blocking the async event loop ([#278](https://github.com/jamiepine/voicebox/issues/278))
-- TypeScript frontend with React Query and Zustand
+- Fixed generation list not updating when a generation completes — switched to `refetchQueries` for reliable cache busting, added SSE error fallback, and page reset on completion ([#231](https://github.com/jamiepine/voicebox/issues/231))
-- Qwen3-TTS for voice cloning
+- Fixed error toasts showing `[object Object]` instead of the actual error message ([#290](https://github.com/jamiepine/voicebox/issues/290))
-- Whisper for transcription
+- Added Whisper model selection (`base`, `small`, `medium`, `large`, `turbo`) and expanded language support to the `/transcribe` endpoint ([#233](https://github.com/jamiepine/voicebox/issues/233))
+- Upgraded CUDA backend build from cu121 to cu126 for RTX 50-series (Blackwell) GPU support ([#289](https://github.com/jamiepine/voicebox/issues/289))
+- Handled client disconnects in SSE and streaming endpoints to suppress `[Errno 32] Broken Pipe` errors ([#248](https://github.com/jamiepine/voicebox/issues/248))
+- Fixed Docker build failure from pip hash mismatch on Qwen3-TTS dependencies ([#286](https://github.com/jamiepine/voicebox/issues/286))
+- Added 50 MB upload size limit with chunked reads to prevent unbounded memory allocation on sample uploads
+- Eliminated redundant double audio decode in sample processing pipeline
+
+### Platform Fixes
+- Replaced `netstat` with `TcpStream` + PowerShell for Windows port detection ([#277](https://github.com/jamiepine/voicebox/pull/277))
+- Fixed Docker frontend build and cleaned up Docker docs
+- Fixed macOS download links to use `.dmg` instead of `.app.tar.gz`
+- Added dynamic download redirect routes to landing site
+
+### Release Tooling
+- Added `draft-release-notes` and `release-bump` agent skills
+- Wired CI release workflow to extract notes from `CHANGELOG.md` for GitHub Releases
+- Backfilled changelog with all historical releases
+
+## [0.2.3] - 2026-03-15
+
+The "it works in dev but not in prod" release. This version fixes a series of PyInstaller bundling issues that prevented model downloading, loading, generation, and progress tracking from working in production builds.
+
+### Model Downloads Now Actually Work
+
+The v0.2.1/v0.2.2 builds could not download or load models that weren't already cached from a dev install. This release fixes the entire chain:
+
+- **Chatterbox, Chatterbox Turbo, and LuxTTS** all download, load, and generate correctly in bundled builds
+- **Real-time download progress** — byte-level progress bars now work in production. The root cause: `huggingface_hub` silently disables tqdm progress bars based on logger level, which prevented our progress tracker from receiving byte updates. We now force-enable the internal counter regardless.
+- **Fixed Python 3.12.0 `code.replace()` bug** — the macOS build was on Python 3.12.0, which has a [known CPython bug](https://github.com/pyinstaller/pyinstaller/issues/7992) that corrupts bytecode when PyInstaller rewrites code objects. This caused `NameError: name 'obj' is not defined` crashes during scipy/torch imports. Upgraded to Python 3.12.13.
+
+### PyInstaller Fixes
+
+- Collect all `inflect` files — `typeguard`'s `@typechecked` decorator calls `inspect.getsource()` at import time, which needs `.py` source files, not just bytecode. Fixes LuxTTS "could not get source code" error.
+- Collect all `perth` files — bundles the pretrained watermark model (`hparams.yaml`, `.pth.tar`) needed by Chatterbox at runtime
+- Collect all `piper_phonemize` files — bundles `espeak-ng-data/` (phoneme tables, language dicts) needed by LuxTTS for text-to-phoneme conversion
+- Set `ESPEAK_DATA_PATH` in frozen builds so the espeak-ng C library finds the bundled data instead of looking at `/usr/share/espeak-ng-data/`
+- Collect all `linacodec` files — fixes `inspect.getsource` error in Vocos codec
+- Collect all `zipvoice` files — fixes source code lookup in LuxTTS voice cloning
+- Copy metadata for `requests`, `transformers`, `huggingface-hub`, `tokenizers`, `safetensors`, `tqdm` — fixes `importlib.metadata` lookups in frozen binary
+- Add hidden imports for `chatterbox`, `chatterbox_turbo`, `luxtts`, `zipvoice` backends
+- Add `multiprocessing.freeze_support()` to fix resource_tracker subprocess crash in frozen binary
+- `--noconsole` now only applied on Windows — macOS/Linux need stdout/stderr for Tauri sidecar log capture
+- Hardened `sys.stdout`/`sys.stderr` devnull redirect to test writability, not just `None` check
+
+### Updater
+
+- Fixed updater artifact generation with `v1Compatible` for `tauri-action` signature files
+- Updated `tauri-action` to v0.6 to fix updater JSON and `.sig` generation
+
+### Other Fixes
+
+- Full traceback logging on all backend model loading errors (was just `str(e)` before)
+
+## [0.2.2] - 2026-03-15
+
+- Fix Chatterbox model support in bundled builds
+- Fix LuxTTS/ZipVoice support in bundled builds
+- Auto-update CUDA binary when app version changes
+- CUDA download progress bar
+- Fix server process staying alive on macOS (SIGHUP handling, watchdog grace period)
+- Hide console window when running CUDA binary on Windows
+
+## [0.2.1] - 2026-03-15
+
+Voicebox v0.1.x was a single-engine voice cloning app built around Qwen3-TTS. v0.2.0 is a ground-up rethink: four TTS engines, 23 languages, paralinguistic emotion controls, a post-processing effects pipeline, unlimited generation length, an async generation queue, and support for every major GPU vendor. Plus Docker.
+
+### New TTS Engines
+
+#### Multi-Engine Architecture
+
+Voicebox now runs **four independent TTS engines** behind a thread-safe per-engine backend registry. Switch engines per-generation from a single dropdown — no restart required.
+
+| Engine                      | Languages | Size    | Key Strengths                                 |
+| --------------------------- | --------- | ------- | --------------------------------------------- |
+| **Qwen3-TTS 1.7B**          | 10        | ~3.5 GB | Highest quality, delivery instructions        |
+| **Qwen3-TTS 0.6B**          | 10        | ~1.2 GB | Lighter, faster variant                       |
+| **LuxTTS**                  | English   | ~300 MB | CPU-friendly, 48 kHz output, 150x realtime    |
+| **Chatterbox Multilingual** | 23        | ~3.2 GB | Broadest language coverage, zero-shot cloning |
+| **Chatterbox Turbo**        | English   | ~1.5 GB | 350M params, low latency, paralinguistic tags |
+
+#### Chatterbox Multilingual — 23 Languages ([#257](https://github.com/jamiepine/voicebox/pull/257))
+
+Zero-shot voice cloning in Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish.
+
+#### LuxTTS — Lightweight English TTS ([#254](https://github.com/jamiepine/voicebox/pull/254))
+
+A fast, CPU-friendly English engine. ~300 MB download, 48 kHz output, runs at 150x realtime on CPU.
+
+#### Chatterbox Turbo — Expressive English ([#258](https://github.com/jamiepine/voicebox/pull/258))
+
+A fast 350M-parameter English model with inline paralinguistic tags.
+
+#### Paralinguistic Tags Autocomplete ([#265](https://github.com/jamiepine/voicebox/pull/265))
+
+Type `/` in the text input with Chatterbox Turbo selected to open an autocomplete for **9 expressive tags**: `[laugh]` `[chuckle]` `[gasp]` `[cough]` `[sigh]` `[groan]` `[sniff]` `[shush]` `[clear throat]`
+
+### Generation
+
+#### Unlimited Generation Length — Auto-Chunking ([#266](https://github.com/jamiepine/voicebox/pull/266))
+
+Long text is now automatically split at sentence boundaries, generated per-chunk, and crossfaded back together. Engine-agnostic.
+
+- Auto-chunking limit slider — 100–5,000 chars (default 800)
+- Crossfade slider — 0–200ms (default 50ms)
+- Max text length raised to 50,000 characters
+- Smart splitting respects abbreviations, CJK punctuation, and `[tags]`
+
+#### Asynchronous Generation Queue ([#269](https://github.com/jamiepine/voicebox/pull/269))
+
+Generation is now fully non-blocking. Serial execution queue prevents GPU contention. Real-time SSE status streaming.
+
+#### Generation Versions
+
+Every generation now supports multiple versions with provenance tracking — original, effects versions, takes, source tracking, version pinning in stories, and favorites.
+
+### Post-Processing Effects ([#271](https://github.com/jamiepine/voicebox/pull/271))
+
+A full audio effects system powered by Spotify's `pedalboard` library: Pitch Shift, Reverb, Delay, Chorus/Flanger, Compressor, Gain, High-Pass Filter, Low-Pass Filter. 4 built-in presets, custom presets, per-profile default effects, and live preview.
 
 ### Platform Support
 
-- macOS (Apple Silicon and Intel)
+- **Windows Support** ([#272](https://github.com/jamiepine/voicebox/pull/272)) — Full Windows support with CUDA GPU detection
-- Windows
+- **Linux** ([#262](https://github.com/jamiepine/voicebox/pull/262)) — AMD ROCm, NVIDIA GBM fix, WebKitGTK mic access (build from source)
-- Linux (AppImage)
+- **NVIDIA CUDA Backend Swap** ([#252](https://github.com/jamiepine/voicebox/pull/252)) — Download and swap in CUDA backend from within the app
+- **Intel Arc (XPU) and DirectML** — PyTorch backend supports Intel Arc and DirectML
+- **Docker + Web Deployment** ([#161](https://github.com/jamiepine/voicebox/pull/161)) — 3-stage build, non-root runtime, health checks
+- **Whisper Turbo** — Added `openai/whisper-large-v3-turbo` as a transcription model option
 
----
+### Model Management ([#268](https://github.com/jamiepine/voicebox/pull/268))
 
-## [Unreleased]
+Per-model unload, custom models directory, model folder migration, download cancel/clear UI ([#238](https://github.com/jamiepine/voicebox/pull/238)), restructured settings UI.
 
-### Fixed
+### Security & Reliability
-- Audio export failing when Tauri save dialog returns object instead of string path
-- OpenAPI client generator script now documents the local backend port and avoids an unused loop variable warning
 
-### Added
+- CORS hardening ([#88](https://github.com/jamiepine/voicebox/pull/88))
-- **justfile** - Comprehensive development workflow automation with commands for setup, development, building, testing, and code quality checks
+- Network access toggle ([#133](https://github.com/jamiepine/voicebox/pull/133))
-  - Cross-platform support (macOS, Linux, Windows)
+- Offline crash fix ([#152](https://github.com/jamiepine/voicebox/pull/152))
-  - Python version detection and compatibility warnings
+- Atomic audio saves ([#263](https://github.com/jamiepine/voicebox/pull/263))
-  - Self-documenting help system with `just --list`
+- Filesystem health endpoint
+- Chatterbox float64 dtype fix ([#264](https://github.com/jamiepine/voicebox/pull/264))
 
-### Changed
+### Accessibility ([#243](https://github.com/jamiepine/voicebox/pull/243))
-- **README** - Updated Quick Start with justfile-based setup instructions
 
-### Removed
+Screen reader support, keyboard navigation, state-aware `aria-label` attributes on all interactive controls.
-- **Makefile** - Replaced by justfile (cross-platform, simpler syntax)
 
----
+### UI Polish
 
-## [Unreleased - Planned]
+- Redesigned landing page ([#274](https://github.com/jamiepine/voicebox/pull/274))
+- Voices tab overhaul with inline inspector
+- Responsive layout improvements
+- Duplicate profile name validation ([#175](https://github.com/jamiepine/voicebox/pull/175))
 
-### Planned
+### Community Contributors
-- Real-time streaming synthesis
-- Conversation mode with multiple speakers
-- Voice effects (pitch shift, reverb, M3GAN-style)
-- Timeline-based audio editor
-- Additional voice models (XTTS, Bark)
-- Voice design from text descriptions
-- Project system for saving sessions
-- Plugin architecture
 
----
+[@haosenwang1018](https://github.com/haosenwang1018), [@Balneario-de-Cofrentes](https://github.com/Balneario-de-Cofrentes), [@ageofalgo](https://github.com/ageofalgo), [@mikeswann](https://github.com/mikeswann), [@rayl15](https://github.com/rayl15), [@mpecanha](https://github.com/mpecanha), [@ways2read](https://github.com/ways2read), [@ieguiguren](https://github.com/ieguiguren), [@Vaibhavee89](https://github.com/Vaibhavee89), [@pandego](https://github.com/pandego), [@luminest-llc](https://github.com/luminest-llc)
 
+## [0.1.13] - 2026-02-23
+
+### Stability and reliability
+
+- [#95](https://github.com/jamiepine/voicebox/pull/95) Fix: selecting 0.6B model still downloads and uses 1.7B
+- [#93](https://github.com/jamiepine/voicebox/pull/93) fix(mlx): bundle native libs and broaden error handling for Apple Silicon
+- [#79](https://github.com/jamiepine/voicebox/pull/79) fix: handle non-ASCII filenames in Content-Disposition headers
+- [#78](https://github.com/jamiepine/voicebox/pull/78) fix: guard getUserMedia call against undefined mediaDevices in non-secure contexts
+- [#77](https://github.com/jamiepine/voicebox/pull/77) fix: await for confirmation before deleting voices and channels
+- [#128](https://github.com/jamiepine/voicebox/pull/128) fix: resolve multiple issues (#96, #119, #111, #108, #121, #125, #127)
+- [#40](https://github.com/jamiepine/voicebox/pull/40) Fix: audio export path resolution
+
+### Build and packaging
+
+- [#122](https://github.com/jamiepine/voicebox/pull/122) fix(web): add @tailwindcss/vite plugin to web config
+- [#126](https://github.com/jamiepine/voicebox/pull/126) Create requirements.txt
+
+### UX and docs
+
+- [#44](https://github.com/jamiepine/voicebox/pull/44) Enhances floating generate box UX
+- [#57](https://github.com/jamiepine/voicebox/pull/57) chore: updates repo URL in README
+- [#146](https://github.com/jamiepine/voicebox/pull/146) Add Spacebot banner to landing page
+- [#1](https://github.com/jamiepine/voicebox/pull/1) Improvements
+
+## [0.1.12] - 2026-01-31
+
+### Model Download UX Overhaul
+
+- Real-time download progress tracking with accurate percentage and speed info
+- No more downloading notifications during generation even when its not downloading
+- Better error handling and status reporting throughout the download process
+
+### Other Improvements
+
+- Enhanced health check endpoint with GPU type information
+- Improved model caching verification
+- More reliable SSE progress updates
+- Actual update notifications — no need to manually check in settings anymore
+
+## [0.1.11] - 2026-01-30
+
+- Fixed transcriptions on MLX
+- Fixed model download progress (finally)
+
+## [0.1.10] - 2026-01-30
+
+### Faster generation on Apple Silicon
+
+Massive speed gains, from around 20s per generation to 2-3s. Added native MLX backend support for Apple Silicon, providing significantly faster TTS and STT generation on M-series macOS machines.
+
+- **MLX Backend** — New backend implementation optimized for Apple Silicon using MLX framework
+- **Dynamic Backend Selection** — Automatically detects platform and selects between MLX (macOS) and PyTorch (other platforms)
+- Refactored TTS and STT logic into modular backend implementations
+- Updated build process to include MLX-specific dependencies for macOS builds
+
+## [0.1.9] - 2026-01-30
+
+### Improved voice profile creation flow
+
+- Voice create drafts: No longer lose work if you close the modal
+- Fixed whisper only transcribing English or Chinese, now has support for all languages
+
+### Improved Stories editor
+
+- Added spacebar for play/pause
+- Timeline now auto-scrolls to follow playhead during playback
+- Fixed misalignment of the items with mouse when picking up
+- Fixed hitbox for selecting an item
+- Fixed playhead jumping forward when pressing play
+
+### Generation box improvements
+
+- Instruct mode no longer wipes prompt text
+- Improved UI cleanliness
+
+### Misc
+
+- Fixed "Model downloading" toast during generation when model is already downloaded
+
+## [0.1.8] - 2026-01-29
+
+### Model Download Timeout Issues
+
+Fixed critical issue where model downloads would fail with "Failed to fetch" errors on Windows. Refactored download endpoints to return immediately and continue downloads in background.
+
+### Cross-Platform Cache Path Issues
+
+Fixed hardcoded `~/.cache/huggingface/hub` paths that don't work on Windows. All cache paths now use `hf_constants.HF_HUB_CACHE` for proper cross-platform support.
+
+### Windows Process Management
+
+- Added `/shutdown` endpoint for graceful server shutdown on Windows
+- Added `gpu_type` field to health check response
+
+## [0.1.7] - 2026-01-29
+
+- Trim and split audio clips in Story Editor
+- Auto-activation of stories in Story Editor with visible playhead
+- Conditional auto-play support in AudioPlayer for better user control
+- Refactored audio loading across HistoryTable, SampleList, and generation forms
+- Audio now only auto-plays when explicitly intended, preventing unexpected playback
+
+## [0.1.6] - 2026-01-29
+
+### Introducing Stories
+
+A full voice editor for composing podcasts and generated conversations.
+
+- **Stories Editor** — Create multi-voice narratives, podcasts, or conversations with a timeline-based editor
+- Compose tracks with different voices
+- Edit and arrange audio segments inline
+- Build generated conversations with multiple participants
+- **Improved Voice Generation UI** — Auto-resizing input, default voice selection, better layout
+- **Track Editor Integration** — Inline track editing within story items
+
+## [0.1.5] - 2026-01-28
+
+Fixed recording length limit at 0:29 to auto stop instead of passing the limit and getting an error, which would cause users to lose their recording.
+
+## [0.1.4] - 2026-01-28
+
+- Audio channel management system
+- Native audio playback handling in AudioPlayer component
+- Refactored ConnectionForm and Checkbox components
+- Improved layout consistency and responsiveness
+- Added safe area constants for better responsive design
+
+## [0.1.3] - 2026-01-27
+
+- Improved the generate textbox
+- Maybe fixed Windows autoupdate restarting entire computer
+
+## [0.1.2] - 2026-01-27
+
+### Audio Capture & Format Conversion
+
+- Added audio format conversion util
+- Enhanced system audio capture on macOS and Windows
+- Improved audio recording hooks
+- Added audio input entitlement for macOS
+- Added audio capture tests
+
+### Update System
+
+- Enhanced auto-updater functionality and update status display
+
+## [0.1.1] - 2026-01-27
+
+### Platform Support
+
+- **macOS Audio Capture** — Native audio capture support for sample creation
+- **Windows Audio Capture** — WASAPI implementation with improved thread safety
+- **Linux Support** — Temporarily removed builds due to runner disk space constraints
+
+### Audio Features
+
+- Play/pause for audio samples across all components
+- Three new sample components: Recording, System capture, Upload with drag-and-drop
+- Audio validation, error handling, and consistent cleanup
+
+### Voice Profile Management
+
+- Profile import with file size validation (100MB limit)
+- Enhanced profile form with new audio sample components
+- Drag-and-drop support for audio file uploads
+
+### Server Management
+
+- Changed default URL from `localhost:8000` to `127.0.0.1:17493`
+- Server reuse logic, "keep server running" preference, orphaned process handling
+
+### Build & Release
+
+- Added `.bumpversion.cfg` for automated version management
+- Enhanced icon generation script for multi-size Windows icons
+
+### Bug Fixes
+
+- Fixed date formatting for timezone-less date strings
+- Fixed getLatestRelease file filtering
+- Improved audio duration metadata on Windows
+
+## [0.1.0] - 2026-01-27
+
+The first public release of Voicebox — an open-source voice synthesis studio powered by Qwen3-TTS.
+
+### Voice Cloning with Qwen3-TTS
+
+- Automatic model download from HuggingFace
+- Multiple model sizes (1.7B and 0.6B)
+- Voice prompt caching for instant regeneration
+- English and Chinese support
+
+### Voice Profile Management
+
+- Create profiles from audio files or record directly in the app
+- Multiple samples per profile for higher quality cloning
+- Import/Export profiles
+- Automatic transcription via Whisper
+
+### Speech Generation
+
+- Simple text-to-speech with profile selection
+- Seed control for reproducible generations
+- Long-form support up to 5,000 characters
+
+### Generation History
+
+- Full history with metadata
+- Search by text content
+- Inline playback and download
+
+### Flexible Deployment
+
+- Local mode with bundled backend
+- Remote mode for GPU servers on your network
+- One-click server setup
+
+### Desktop Experience
+
+- Built with Tauri v2 (Rust) — native performance, not Electron
+- Cross-platform: macOS and Windows
+- No Python installation required
+
+### Tech Stack
+
+Tauri v2, React, TypeScript, Tailwind CSS, FastAPI, Qwen3-TTS, Whisper, SQLite
+
+[Unreleased]: https://github.com/jamiepine/voicebox/compare/v0.2.3...HEAD
+[0.2.3]: https://github.com/jamiepine/voicebox/compare/v0.2.2...v0.2.3
+[0.2.2]: https://github.com/jamiepine/voicebox/compare/v0.2.1...v0.2.2
+[0.2.1]: https://github.com/jamiepine/voicebox/compare/v0.1.13...v0.2.1
+[0.1.13]: https://github.com/jamiepine/voicebox/compare/v0.1.12...v0.1.13
+[0.1.12]: https://github.com/jamiepine/voicebox/compare/v0.1.11...v0.1.12
+[0.1.11]: https://github.com/jamiepine/voicebox/compare/v0.1.10...v0.1.11
+[0.1.10]: https://github.com/jamiepine/voicebox/compare/v0.1.9...v0.1.10
+[0.1.9]: https://github.com/jamiepine/voicebox/compare/v0.1.8...v0.1.9
+[0.1.8]: https://github.com/jamiepine/voicebox/compare/v0.1.7...v0.1.8
+[0.1.7]: https://github.com/jamiepine/voicebox/compare/v0.1.6...v0.1.7
+[0.1.6]: https://github.com/jamiepine/voicebox/compare/v0.1.5...v0.1.6
+[0.1.5]: https://github.com/jamiepine/voicebox/compare/v0.1.4...v0.1.5
+[0.1.4]: https://github.com/jamiepine/voicebox/compare/v0.1.3...v0.1.4
+[0.1.3]: https://github.com/jamiepine/voicebox/compare/v0.1.2...v0.1.3
+[0.1.2]: https://github.com/jamiepine/voicebox/compare/v0.1.1...v0.1.2
+[0.1.1]: https://github.com/jamiepine/voicebox/compare/v0.1.0...v0.1.1
 [0.1.0]: https://github.com/jamiepine/voicebox/releases/tag/v0.1.0

Dockerfile

@@ -31,8 +31,12 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
     build-essential \
     && rm -rf /var/lib/apt/lists/*
 
+RUN pip install --no-cache-dir --upgrade pip
+
 COPY backend/requirements.txt .
 RUN pip install --no-cache-dir --prefix=/install -r requirements.txt
+RUN pip install --no-cache-dir --prefix=/install --no-deps chatterbox-tts
+RUN pip install --no-cache-dir --prefix=/install --no-deps hume-tada
 RUN pip install --no-cache-dir --prefix=/install \
     git+https://github.com/QwenLM/Qwen3-TTS.git
 

PATCH_NOTES.md

@@ -1,58 +0,0 @@
-# Voicebox Offline Mode Fix
-
-## Problem
-Voicebox crashes when generating speech if HuggingFace is unreachable, even when models are fully cached locally.
-
-**Root Cause:**
-- Voicebox downloads `mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16` (MLX optimized version)
-- But `mlx_audio.tts.load()` tries to fetch `config.json` from original repo `Qwen/Qwen3-TTS-12Hz-1.7B-Base`
-- This network request fails → server crashes with `RemoteDisconnected`
-
-**Related Issues:**
-- Issue #150: "Internet connection required, even though models are downloaded?"
-- Issue #151: "API Stability Issues: Model Loading Hangs and Server Crashes"
-
-## Solution
-Two-part fix:
-
-### 1. Monkey-patch huggingface_hub (`backend/utils/hf_offline_patch.py`)
-- Intercepts cache lookup functions
-- Forces offline mode early (before mlx_audio imports)
-- Adds debug logging for cache hits/misses
-
-### 2. Symlink original repo to MLX version (`ensure_original_qwen_config_cached()`)
-- When original `Qwen/Qwen3-TTS-12Hz-1.7B-Base` cache doesn't exist
-- But MLX `mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16` does exist
-- Creates a symlink so cache lookups succeed
-
-## Files Changed
-- `backend/backends/mlx_backend.py` - Added patch imports at top
-- `backend/utils/hf_offline_patch.py` - New patch module
-
-## Testing
-To test this fix:
-1. Build Voicebox from source: `just build`
-2. Disconnect from internet
-3. Try generating speech
-4. Should work without network requests
-
-## Build Instructions
-
-```bash
-# Install dependencies
-just setup
-
-# Build the app
-just build
-
-# Or build just the server
-just build-server
-```
-
-## Notes
-- The patch is applied automatically when `mlx_backend.py` is imported
-- Set `VOICEBOX_OFFLINE_PATCH=0` to disable the patch
-- The symlink approach works because the config.json is compatible between versions
-
----
-*Patch contributed by community*

README.md

@@ -27,10 +27,10 @@
 
 <p align="center">
   <a href="https://voicebox.sh">voicebox.sh</a> •
+  <a href="https://docs.voicebox.sh">Docs</a> •
   <a href="#download">Download</a> •
   <a href="#features">Features</a> •
-  <a href="#api">API</a> •
+  <a href="#api">API</a>
-  <a href="#roadmap">Roadmap</a>
 </p>
 
 <br/>
@@ -59,10 +59,10 @@
 
 ## What is Voicebox?
 
-Voicebox is a **local-first voice cloning studio** — a free and open-source alternative to ElevenLabs. Clone voices from a few seconds of audio, generate speech in 23 languages across 4 TTS engines, apply post-processing effects, and compose multi-voice projects with a timeline editor.
+Voicebox is a **local-first voice cloning studio** — a free and open-source alternative to ElevenLabs. Clone voices from a few seconds of audio, generate speech in 23 languages across 5 TTS engines, apply post-processing effects, and compose multi-voice projects with a timeline editor.
 
 - **Complete privacy** — models and voice data stay on your machine
-- **4 TTS engines** — Qwen3-TTS, LuxTTS, Chatterbox Multilingual, and Chatterbox Turbo
+- **5 TTS engines** — Qwen3-TTS, LuxTTS, Chatterbox Multilingual, Chatterbox Turbo, and HumeAI TADA
 - **23 languages** — from English to Arabic, Japanese, Hindi, Swahili, and more
 - **Post-processing effects** — pitch shift, reverb, delay, chorus, compression, and filters
 - **Expressive speech** — paralinguistic tags like `[laugh]`, `[sigh]`, `[gasp]` via Chatterbox Turbo
@@ -76,12 +76,12 @@ Voicebox is a **local-first voice cloning studio** — a free and open-source al
 
 ## Download
 
-| Platform | Download |
+| Platform              | Download                                               |
-|----------|----------|
+| --------------------- | ------------------------------------------------------ |
-| macOS (Apple Silicon) | [Download DMG](https://voicebox.sh/download/mac-arm) |
+| macOS (Apple Silicon) | [Download DMG](https://voicebox.sh/download/mac-arm)   |
-| macOS (Intel) | [Download DMG](https://voicebox.sh/download/mac-intel) |
+| macOS (Intel)         | [Download DMG](https://voicebox.sh/download/mac-intel) |
-| Windows | [Download MSI](https://voicebox.sh/download/windows) |
+| Windows               | [Download MSI](https://voicebox.sh/download/windows)   |
-| Docker | `docker compose up` |
+| Docker                | `docker compose up`                                    |
 
 > **[View all binaries →](https://github.com/jamiepine/voicebox/releases/latest)**
 
@@ -93,14 +93,15 @@ Voicebox is a **local-first voice cloning studio** — a free and open-source al
 
 ### Multi-Engine Voice Cloning
 
-Four TTS engines with different strengths, switchable per-generation:
+Five TTS engines with different strengths, switchable per-generation:
 
-| Engine | Languages | Strengths |
+| Engine                      | Languages | Strengths                                                                                                                                |
-|--------|-----------|-----------|
+| --------------------------- | --------- | ---------------------------------------------------------------------------------------------------------------------------------------- |
-| **Qwen3-TTS** (0.6B / 1.7B) | 10 | High-quality multilingual cloning, delivery instructions ("speak slowly", "whisper") |
+| **Qwen3-TTS** (0.6B / 1.7B) | 10        | High-quality multilingual cloning, delivery instructions ("speak slowly", "whisper")                                                     |
-| **LuxTTS** | English | Lightweight (~1GB VRAM), 48kHz output, 150x realtime on CPU |
+| **LuxTTS**                  | English   | Lightweight (~1GB VRAM), 48kHz output, 150x realtime on CPU                                                                              |
-| **Chatterbox Multilingual** | 23 | Broadest language coverage — Arabic, Danish, Finnish, Greek, Hebrew, Hindi, Malay, Norwegian, Polish, Swahili, Swedish, Turkish and more |
+| **Chatterbox Multilingual** | 23        | Broadest language coverage — Arabic, Danish, Finnish, Greek, Hebrew, Hindi, Malay, Norwegian, Polish, Swahili, Swedish, Turkish and more |
-| **Chatterbox Turbo** | English | Fast 350M model with paralinguistic emotion/sound tags |
+| **Chatterbox Turbo**        | English   | Fast 350M model with paralinguistic emotion/sound tags                                                                                   |
+| **TADA** (1B / 3B)          | 10        | HumeAI speech-language model — 700s+ coherent audio, text-acoustic dual alignment                                                        |
 
 ### Emotions & Paralinguistic Tags
 
@@ -112,16 +113,16 @@ Type `/` in the text input to insert expressive tags that the model synthesizes
 
 8 audio effects powered by Spotify's `pedalboard` library. Apply after generation, preview in real time, build reusable presets.
 
-| Effect | Description |
+| Effect           | Description                                   |
-|--------|-------------|
+| ---------------- | --------------------------------------------- |
-| Pitch Shift | Up or down by up to 12 semitones |
+| Pitch Shift      | Up or down by up to 12 semitones              |
-| Reverb | Configurable room size, damping, wet/dry mix |
+| Reverb           | Configurable room size, damping, wet/dry mix  |
-| Delay | Echo with adjustable time, feedback, and mix |
+| Delay            | Echo with adjustable time, feedback, and mix  |
 | Chorus / Flanger | Modulated delay for metallic or lush textures |
-| Compressor | Dynamic range compression |
+| Compressor       | Dynamic range compression                     |
-| Gain | Volume adjustment (-40 to +40 dB) |
+| Gain             | Volume adjustment (-40 to +40 dB)             |
-| High-Pass Filter | Remove low frequencies |
+| High-Pass Filter | Remove low frequencies                        |
-| Low-Pass Filter | Remove high frequencies |
+| Low-Pass Filter  | Remove high frequencies                       |
 
 Ships with 4 built-in presets (Robotic, Radio, Echo Chamber, Deep Voice) and supports custom presets. Effects can be assigned per-profile as defaults.
 
@@ -186,14 +187,14 @@ Multi-voice timeline editor for conversations, podcasts, and narratives.
 
 ### GPU Support
 
-| Platform | Backend | Notes |
+| Platform                 | Backend        | Notes                                          |
-|----------|---------|-------|
+| ------------------------ | -------------- | ---------------------------------------------- |
-| macOS (Apple Silicon) | MLX (Metal) | 4-5x faster via Neural Engine |
+| macOS (Apple Silicon)    | MLX (Metal)    | 4-5x faster via Neural Engine                  |
 | Windows / Linux (NVIDIA) | PyTorch (CUDA) | Auto-downloads CUDA binary from within the app |
-| Linux (AMD) | PyTorch (ROCm) | Auto-configures HSA_OVERRIDE_GFX_VERSION |
+| Linux (AMD)              | PyTorch (ROCm) | Auto-configures HSA_OVERRIDE_GFX_VERSION       |
-| Windows (any GPU) | DirectML | Universal Windows GPU support |
+| Windows (any GPU)        | DirectML       | Universal Windows GPU support                  |
-| Intel Arc | IPEX/XPU | Intel discrete GPU acceleration |
+| Intel Arc                | IPEX/XPU       | Intel discrete GPU acceleration                |
-| Any | CPU | Works everywhere, just slower |
+| Any                      | CPU            | Works everywhere, just slower                  |
 
 ---
 
@@ -224,30 +225,30 @@ Full API documentation available at `http://localhost:17493/docs`.
 
 ## Tech Stack
 
-| Layer | Technology |
+| Layer         | Technology                                        |
-|-------|------------|
+| ------------- | ------------------------------------------------- |
-| Desktop App | Tauri (Rust) |
+| Desktop App   | Tauri (Rust)                                      |
-| Frontend | React, TypeScript, Tailwind CSS |
+| Frontend      | React, TypeScript, Tailwind CSS                   |
-| State | Zustand, React Query |
+| State         | Zustand, React Query                              |
-| Backend | FastAPI (Python) |
+| Backend       | FastAPI (Python)                                  |
-| TTS Engines | Qwen3-TTS, LuxTTS, Chatterbox, Chatterbox Turbo |
+| TTS Engines   | Qwen3-TTS, LuxTTS, Chatterbox, Chatterbox Turbo, TADA |
-| Effects | Pedalboard (Spotify) |
+| Effects       | Pedalboard (Spotify)                              |
-| Transcription | Whisper / Whisper Turbo (PyTorch or MLX) |
+| Transcription | Whisper / Whisper Turbo (PyTorch or MLX)          |
-| Inference | MLX (Apple Silicon) / PyTorch (CUDA/ROCm/XPU/CPU) |
+| Inference     | MLX (Apple Silicon) / PyTorch (CUDA/ROCm/XPU/CPU) |
-| Database | SQLite |
+| Database      | SQLite                                            |
-| Audio | WaveSurfer.js, librosa |
+| Audio         | WaveSurfer.js, librosa                            |
 
 ---
 
 ## Roadmap
 
-| Feature | Description |
+| Feature                 | Description                                    |
-|---------|-------------|
+| ----------------------- | ---------------------------------------------- |
-| **Real-time Streaming** | Stream audio as it generates, word by word |
+| **Real-time Streaming** | Stream audio as it generates, word by word     |
-| **Voice Design** | Create new voices from text descriptions |
+| **Voice Design**        | Create new voices from text descriptions       |
-| **More Models** | XTTS, Bark, and other open-source voice models |
+| **More Models**         | XTTS, Bark, and other open-source voice models  |
-| **Plugin Architecture** | Extend with custom models and effects |
+| **Plugin Architecture** | Extend with custom models and effects          |
-| **Mobile Companion** | Control Voicebox from your phone |
+| **Mobile Companion**    | Control Voicebox from your phone               |
 
 ---
 
@@ -276,6 +277,12 @@ just build          # Build CPU server binary + Tauri app
 just build-local    # (Windows) Build CPU + CUDA server binaries + Tauri app
 ```
 
+### Adding New Voice Models
+
+The multi-engine architecture makes adding new TTS engines straightforward. A [step-by-step guide](docs/content/docs/developer/tts-engines.mdx) covers the full process: dependency research, backend protocol implementation, frontend wiring, and PyInstaller bundling.
+
+The guide is optimized for AI coding agents. An [agent skill](.agents/skills/add-tts-engine/SKILL.md) can pick up a model name and handle the entire integration autonomously — you just test the build locally.
+
 ### Project Structure
 
 ```

app/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@voicebox/app",
-  "version": "0.2.3",
+  "version": "0.3.1",
   "private": true,
   "type": "module",
   "scripts": {

app/plugins/changelog.ts

@@ -0,0 +1,23 @@
+import { readFileSync } from 'node:fs';
+import path from 'node:path';
+import type { Plugin } from 'vite';
+
+/** Vite plugin that exposes CHANGELOG.md as `virtual:changelog`. */
+export function changelogPlugin(repoRoot: string): Plugin {
+  const virtualId = 'virtual:changelog';
+  const resolvedId = '\0' + virtualId;
+  const changelogPath = path.resolve(repoRoot, 'CHANGELOG.md');
+
+  return {
+    name: 'changelog',
+    resolveId(id) {
+      if (id === virtualId) return resolvedId;
+    },
+    load(id) {
+      if (id === resolvedId) {
+        const raw = readFileSync(changelogPath, 'utf-8');
+        return `export default ${JSON.stringify(raw)};`;
+      }
+    },
+  };
+}

app/src/App.tsx

@@ -8,6 +8,7 @@ import { TOP_SAFE_AREA_PADDING } from '@/lib/constants/ui';
 import { cn } from '@/lib/utils/cn';
 import { usePlatform } from '@/platform/PlatformContext';
 import { router } from '@/router';
+import { useLogStore } from '@/stores/logStore';
 import { useServerStore } from '@/stores/serverStore';
 
 const LOADING_MESSAGES = [
@@ -63,6 +64,14 @@ function App() {
     // eslint-disable-next-line react-hooks/exhaustive-deps
   }, [platform.lifecycle]);
 
+  // Subscribe to server logs
+  useEffect(() => {
+    const unsubscribe = platform.lifecycle.subscribeToServerLogs((entry) => {
+      useLogStore.getState().addEntry(entry);
+    });
+    return unsubscribe;
+  }, [platform.lifecycle]);
+
   // Setup window close handler and auto-start server when running in Tauri (production only)
   useEffect(() => {
     if (!platform.metadata.isTauri) {

app/src/components/AudioPlayer/AudioPlayer.tsx

@@ -17,7 +17,6 @@ export function AudioPlayer() {
     audioUrl,
     audioId,
     profileId,
-    title,
     isPlaying,
     currentTime,
     duration,
@@ -63,7 +62,7 @@ export function AudioPlayer() {
     );
 
     return shouldUseNative;
-  }, [profileChannels, channels, profileId]);
+  }, [profileChannels, channels, platform.metadata.isTauri]);
 
   const waveformRef = useRef<HTMLDivElement>(null);
   const wavesurferRef = useRef<WaveSurfer | null>(null);
@@ -73,31 +72,21 @@ export function AudioPlayer() {
   const isUsingNativePlaybackRef = useRef(false);
   const [isLoading, setIsLoading] = useState(false);
   const [error, setError] = useState<string | null>(null);
+  const [wsReady, setWsReady] = useState(false);
 
-  // Initialize WaveSurfer (only when audioUrl exists and container is ready)
+  // Create WaveSurfer once when the player becomes visible (audioUrl is set).
+  // This instance is reused for all subsequent audio loads - never destroyed until unmount.
   useEffect(() => {
-    // Don't initialize if no audioUrl or already initialized
+    if (!audioUrl) return;
-    if (!audioUrl) {
+    if (wavesurferRef.current) return; // already created
-      return;
-    }
 
-    if (wavesurferRef.current) {
-      debug.log('WaveSurfer already initialized, skipping');
-      return;
-    }
-
-    debug.log('Creating NEW WaveSurfer instance');
-
-    // Wait for container to be properly rendered
     const initWaveSurfer = () => {
       const container = waveformRef.current;
       if (!container) {
-        // Container not ready yet, retry
         setTimeout(initWaveSurfer, 50);
         return;
       }
 
-      // Check if container has dimensions and is visible
       const rect = container.getBoundingClientRect();
       const style = window.getComputedStyle(container);
       const isVisible =
@@ -107,412 +96,221 @@ export function AudioPlayer() {
         style.visibility !== 'hidden';
 
       if (!isVisible) {
-        // Retry after a short delay
         setTimeout(initWaveSurfer, 50);
         return;
       }
 
-      debug.log('Initializing WaveSurfer...', {
+      debug.log('Creating WaveSurfer instance', {
-        container,
         width: rect.width,
         height: rect.height,
       });
 
       try {
-        // Get computed CSS variable values
         const root = document.documentElement;
         const getCSSVar = (varName: string) => {
           const value = getComputedStyle(root).getPropertyValue(varName).trim();
           return value ? `hsl(${value})` : '';
         };
 
-        const waveColor = getCSSVar('--muted');
-        const progressColor = getCSSVar('--accent');
-        const cursorColor = getCSSVar('--accent');
-
         const wavesurfer = WaveSurfer.create({
-          container: container,
+          container,
-          waveColor: waveColor,
+          waveColor: getCSSVar('--muted'),
-          progressColor: progressColor,
+          progressColor: getCSSVar('--accent'),
-          cursorColor: cursorColor,
+          cursorColor: getCSSVar('--accent'),
+          cursorWidth: 3,
           barWidth: 2,
           barRadius: 2,
           height: 80,
           normalize: true,
-          // Use MediaElement backend (default). Unlike the WebAudio backend,
+          interact: true,
-          // MediaElement uses a standard <audio> element for playback which
+          dragToSeek: { debounceTime: 0 },
-          // benefits from the browser/webview's built-in audio session recovery.
+          mediaControls: false,
-          // This prevents audio loss when another app steals audio output or
+          backend: 'WebAudio',
-          // the system audio session is interrupted.
-          interact: true, // Enable interaction (click to seek)
-          mediaControls: false, // Don't show native controls
         });
 
-        wavesurferRef.current = wavesurfer;
+        // Wire up event handlers (these persist for the lifetime of the instance)
-        debug.log('WaveSurfer created successfully');
+        wavesurfer.on('timeupdate', (time) => {
-      } catch (error) {
+          const dur = usePlayerStore.getState().duration;
-        debug.error('Failed to create WaveSurfer:', error);
+          if (dur > 0 && time >= dur) {
-        setError(
+            setCurrentTime(dur);
-          `Failed to initialize waveform: ${error instanceof Error ? error.message : String(error)}`,
+            const loop = usePlayerStore.getState().isLooping;
-        );
+            if (loop) {
-        return;
+              wavesurfer.seekTo(0);
-      }
+              wavesurfer.play().catch((err) => debug.error('Loop play failed:', err));
+            } else {
+              wavesurfer.pause();
+              setIsPlaying(false);
+            }
+            return;
+          }
+          setCurrentTime(time);
+        });
 
-      const wavesurfer = wavesurferRef.current;
+        wavesurfer.on('ready', () => {
-      if (!wavesurfer) return;
+          const dur = wavesurfer.getDuration();
+          setDuration(dur);
+          loadingRef.current = false;
+          setIsLoading(false);
+          setError(null);
+          debug.log('Audio ready, duration:', dur);
 
-      // Update store when time changes, stop if past duration
+          wavesurfer.setVolume(usePlayerStore.getState().volume);
-      wavesurfer.on('timeupdate', (time) => {
+          wavesurfer.setMuted(false);
-        const dur = usePlayerStore.getState().duration;
+
-        if (dur > 0 && time >= dur) {
+          // Auto-play if the flag is set (story mode advance or explicit play)
-          setCurrentTime(dur);
+          const shouldAutoPlayNow = usePlayerStore.getState().shouldAutoPlay;
+          if (shouldAutoPlayNow) {
+            usePlayerStore.getState().clearAutoPlayFlag();
+            wavesurfer.play().catch((err) => {
+              debug.error('Failed to autoplay:', err);
+            });
+          } else {
+            debug.log('Skipping auto-play - shouldAutoPlay is false');
+          }
+        });
+
+        wavesurfer.on('play', () => setIsPlaying(true));
+        wavesurfer.on('pause', () => {
+          setIsPlaying(false);
+          setCurrentTime(wavesurfer.getCurrentTime());
+        });
+
+        wavesurfer.on('seeking', (time) => setCurrentTime(time));
+
+        // Mute audio during drag-to-seek to prevent popping from the WebAudio
+        // backend's hard stop/start cycle on each seek. Unmute with a short
+        // fade-in when the drag ends.
+        const seekMedia = wavesurfer.getMediaElement() as any;
+        const seekGain: GainNode | null = seekMedia?.getGainNode?.() ?? null;
+        if (seekGain) {
+          const ctx = seekGain.context as AudioContext;
+          wavesurfer.on('dragstart', () => {
+            seekGain.gain.cancelScheduledValues(ctx.currentTime);
+            seekGain.gain.setTargetAtTime(0, ctx.currentTime, 0.002);
+          });
+          wavesurfer.on('dragend', () => {
+            seekGain.gain.cancelScheduledValues(ctx.currentTime);
+            seekGain.gain.setTargetAtTime(1, ctx.currentTime, 0.01);
+          });
+        }
+        wavesurfer.on('finish', () => {
           const loop = usePlayerStore.getState().isLooping;
           if (loop) {
             wavesurfer.seekTo(0);
-            wavesurfer.play();
+            wavesurfer.play().catch((err) => debug.error('Loop play failed:', err));
           } else {
-            wavesurfer.pause();
             setIsPlaying(false);
+            const onFinish = usePlayerStore.getState().onFinish;
+            if (onFinish) onFinish();
           }
-          return;
-        }
-        setCurrentTime(time);
-      });
-
-      // Update store when duration is loaded
-      wavesurfer.on('ready', async () => {
-        const dur = wavesurfer.getDuration();
-        setDuration(dur);
-        loadingRef.current = false;
-        setIsLoading(false);
-        setError(null);
-        debug.log('Audio ready, duration:', dur);
-        debug.log('Waveform should be visible now');
-
-        // Ensure volume is set
-        const currentVolume = usePlayerStore.getState().volume;
-        wavesurfer.setVolume(currentVolume);
-
-        // Auto-play when ready - check if we should use native playback
-        // Get current values from the store and queries at runtime (not captured closure values)
-        const currentAudioUrl = usePlayerStore.getState().audioUrl;
-        const currentProfileId = usePlayerStore.getState().profileId;
-
-        debug.log('Auto-play check - capturing runtime values...');
-
-        // Fetch profile channels at runtime (not using captured value)
-        let runtimeProfileChannels = null;
-        let runtimeChannels = null;
-
-        if (platform.metadata.isTauri && currentProfileId) {
-          try {
-            runtimeProfileChannels = await apiClient.getProfileChannels(currentProfileId);
-            debug.log('Runtime profileChannels:', runtimeProfileChannels);
-
-            if (runtimeProfileChannels && runtimeProfileChannels.channel_ids.length > 0) {
-              runtimeChannels = await apiClient.listChannels();
-              debug.log('Runtime channels:', runtimeChannels);
-            }
-          } catch (error) {
-            debug.error('Failed to fetch runtime channel data:', error);
-          }
-        }
-
-        debug.log('Auto-play check:', {
-          isTauri: platform.metadata.isTauri,
-          currentAudioUrl,
-          currentProfileId,
-          hasProfileChannels: !!runtimeProfileChannels,
-          hasChannels: !!runtimeChannels,
         });
 
-        if (
+        wavesurfer.on('error', (err) => {
-          platform.metadata.isTauri &&
+          debug.error('WaveSurfer error:', err);
-          currentAudioUrl &&
-          currentProfileId &&
-          runtimeProfileChannels &&
-          runtimeChannels
-        ) {
-          debug.log('Attempting native audio playback...');
-
-          // Stop any existing native playback first
-          if (isUsingNativePlaybackRef.current) {
-            try {
-              platform.audio.stopPlayback();
-              debug.log('Stopped existing native playback before starting new one');
-            } catch (error) {
-              debug.error('Failed to stop existing playback:', error);
-            }
-          }
-
-          try {
-            // Collect all device IDs from assigned channels
-            const assignedChannels = runtimeChannels.filter((ch: any) =>
-              runtimeProfileChannels.channel_ids.includes(ch.id),
-            );
-            debug.log('Assigned channels for playback:', assignedChannels);
-
-            // Check if any assigned channel has non-default devices
-            const shouldUseNative = assignedChannels.some(
-              (ch: any) => ch.device_ids.length > 0 && !ch.is_default,
-            );
-            debug.log('Should use native playback:', shouldUseNative);
-
-            if (!shouldUseNative) {
-              debug.log('No custom devices assigned, using standard playback');
-              isUsingNativePlaybackRef.current = false;
-            } else {
-              const deviceIds = assignedChannels.flatMap((ch: any) => ch.device_ids);
-              debug.log('Device IDs to play to:', deviceIds);
-
-              if (deviceIds.length > 0) {
-                debug.log('Fetching audio data from:', currentAudioUrl);
-                // Fetch audio data
-                const response = await fetch(currentAudioUrl);
-                const audioData = new Uint8Array(await response.arrayBuffer());
-                debug.log('Audio data size:', audioData.length);
-
-                // Play via native audio
-                debug.log('Invoking play_audio_to_devices...');
-                try {
-                  await platform.audio.playToDevices(audioData, deviceIds);
-                  debug.log('play_audio_to_devices completed successfully');
-
-                  // Mark that we're using native playback
-                  isUsingNativePlaybackRef.current = true;
-
-                  // Mute WaveSurfer's audio output — native handles the actual sound
-                  // Keep WaveSurfer running for waveform visualization
-                  wavesurfer.setVolume(0);
-                  wavesurfer.setMuted(true);
-
-                  // Start WaveSurfer playback for visualization (muted)
-                  wavesurfer.play().catch((error) => {
-                    debug.error('Failed to start WaveSurfer visualization:', error);
-                  });
-
-                  setIsPlaying(true);
-                  debug.log('Auto-playing via native audio routing - SUCCESS');
-                  return;
-                } catch (invokeError) {
-                  debug.error('play_audio_to_devices invoke failed:', invokeError);
-                  throw invokeError;
-                }
-              } else {
-                debug.log('No device IDs found, falling back to WaveSurfer');
-              }
-            }
-          } catch (error) {
-            debug.error(
-              'Native playback failed during auto-play, falling back to WaveSurfer:',
-              error,
-            );
-            isUsingNativePlaybackRef.current = false;
-            // Fall through to WaveSurfer playback
-          }
-        }
-
-        // Standard playback path — ensure WaveSurfer is unmuted
-        if (!isUsingNativePlaybackRef.current) {
-          wavesurfer.setMuted(false);
-          wavesurfer.setVolume(usePlayerStore.getState().volume);
-        }
-
-        // Only auto-play if shouldAutoPlay flag is set (user explicitly clicked to play)
-        const shouldAutoPlayNow = usePlayerStore.getState().shouldAutoPlay;
-        if (shouldAutoPlayNow) {
-          // Clear the flag first
-          usePlayerStore.getState().clearAutoPlayFlag();
-
-          // Use a small delay to ensure audio element is fully ready
-          setTimeout(() => {
-            wavesurfer.play().catch((error) => {
-              debug.error('Failed to autoplay:', error);
-              // Don't show error for autoplay failures (browser restrictions)
-            });
-          }, 100);
-        } else {
-          debug.log('Skipping auto-play - shouldAutoPlay is false');
-        }
-      });
-
-      // Handle play/pause
-      wavesurfer.on('play', () => {
-        setIsPlaying(true);
-      });
-      wavesurfer.on('pause', () => setIsPlaying(false));
-      wavesurfer.on('finish', () => {
-        // Check loop state from store
-        const loop = usePlayerStore.getState().isLooping;
-        if (loop) {
-          wavesurfer.seekTo(0);
-          wavesurfer.play();
-        } else {
-          setIsPlaying(false);
-          // Trigger finish callback if set
-          const onFinish = usePlayerStore.getState().onFinish;
-          if (onFinish) {
-            onFinish();
-          }
-        }
-      });
-
-      // Handle errors
-      wavesurfer.on('error', (error) => {
-        debug.error('WaveSurfer error:', error);
-        setIsLoading(false);
-        setError(`Audio error: ${error instanceof Error ? error.message : String(error)}`);
-      });
-
-      // Handle loading
-      wavesurfer.on('loading', (percent) => {
-        setIsLoading(true);
-        if (percent === 100) {
           setIsLoading(false);
-        }
+          setError(`Audio error: ${err instanceof Error ? err.message : String(err)}`);
-      });
+        });
 
-      // Load audio immediately if audioUrl is already set
+        wavesurfer.on('loading', (percent) => {
-      if (audioUrl) {
+          setIsLoading(true);
-        debug.log('WaveSurfer ready, loading audio:', audioUrl);
+          if (percent === 100) setIsLoading(false);
-        loadingRef.current = true;
+        });
-        setIsLoading(true);
+
-        // Stop any current playback before loading new audio
+        wavesurferRef.current = wavesurfer;
-        if (wavesurfer.isPlaying()) {
+        setWsReady(true);
-          wavesurfer.pause();
+        debug.log('WaveSurfer created successfully');
-        }
+      } catch (err) {
-        wavesurfer
+        debug.error('Failed to create WaveSurfer:', err);
-          .load(audioUrl)
+        setError(
-          .then(() => {
+          `Failed to initialize waveform: ${err instanceof Error ? err.message : String(err)}`,
-            debug.log('Audio loaded into WaveSurfer');
+        );
-            loadingRef.current = false;
-          })
-          .catch((error) => {
-            debug.error('Failed to load audio into WaveSurfer:', error);
-            loadingRef.current = false;
-            setIsLoading(false);
-            setError(
-              `Failed to load audio: ${error instanceof Error ? error.message : String(error)}`,
-            );
-          });
       }
     };
 
-    // Use double requestAnimationFrame to ensure DOM is fully rendered
+    let rafId: number;
-    let rafId1: number;
+    rafId = requestAnimationFrame(() => {
-    let rafId2: number;
+      initWaveSurfer();
-    let timeoutId: number | null = null;
-
-    rafId1 = requestAnimationFrame(() => {
-      rafId2 = requestAnimationFrame(() => {
-        // Add a small delay to ensure container is fully laid out
-        timeoutId = setTimeout(() => {
-          initWaveSurfer();
-        }, 10);
-      });
     });
 
     return () => {
-      debug.log('Cleaning up WaveSurfer initialization effect');
+      cancelAnimationFrame(rafId);
-      if (rafId1) cancelAnimationFrame(rafId1);
+    };
-      if (rafId2) cancelAnimationFrame(rafId2);
+    // Only run on mount-like conditions. audioUrl is here so we create the instance
-      if (timeoutId) clearTimeout(timeoutId);
+    // when the player first appears, but we guard against re-creation above.
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [audioUrl, setIsPlaying, setDuration, setCurrentTime]);
+
+  // Destroy WaveSurfer only on unmount
+  useEffect(() => {
+    return () => {
       if (wavesurferRef.current) {
-        debug.log('Destroying WaveSurfer instance');
+        debug.log('Destroying WaveSurfer instance (unmount)');
         try {
           wavesurferRef.current.destroy();
-        } catch (error) {
+        } catch (err) {
-          debug.error('Error destroying WaveSurfer:', error);
+          debug.error('Error destroying WaveSurfer:', err);
         }
         wavesurferRef.current = null;
+        setWsReady(false);
       }
     };
-  }, [audioUrl, setIsPlaying, setCurrentTime, setDuration]);
+  }, []);
 
-  // Load audio when URL changes (only if WaveSurfer is already initialized)
+  // Load audio when URL changes (reuses the existing WaveSurfer instance)
   useEffect(() => {
     const wavesurfer = wavesurferRef.current;
+    if (!wavesurfer || !wsReady) return;
 
-    if (!audioUrl || !wavesurfer) {
+    if (!audioUrl) {
-      // Reset state when no audio or WaveSurfer not ready
+      // No audio - pause and reset
-      if (!audioUrl && wavesurfer) {
+      wavesurfer.pause();
-        wavesurfer.pause();
+      wavesurfer.seekTo(0);
-        wavesurfer.seekTo(0);
+      loadingRef.current = false;
-        loadingRef.current = false;
+      setIsLoading(false);
-        setIsLoading(false);
+      setDuration(0);
-        setDuration(0);
+      setCurrentTime(0);
-        setCurrentTime(0);
+      setError(null);
-        setError(null);
+      isUsingNativePlaybackRef.current = false;
-        // Reset native playback flag
-        isUsingNativePlaybackRef.current = false;
-      }
       return;
     }
 
-    // Stop native playback if it was active
+    // Reset native playback state
-    if (isUsingNativePlaybackRef.current && platform.metadata.isTauri) {
-      try {
-        platform.audio.stopPlayback();
-        debug.log('Stopped native audio playback');
-      } catch (error) {
-        debug.error('Failed to stop native playback:', error);
-      }
-    }
-
-    // Reset native playback flag when loading new audio
-    // Unmute WaveSurfer if it was muted for native playback
-    if (isUsingNativePlaybackRef.current) {
-      wavesurfer.setMuted(false);
-      wavesurfer.setVolume(usePlayerStore.getState().volume);
-    }
     isUsingNativePlaybackRef.current = false;
+    wavesurfer.setMuted(false);
+    wavesurfer.setVolume(usePlayerStore.getState().volume);
 
-    // CRITICAL: Force stop any current playback and cancel any pending loads
+    // Stop current playback and reset position before loading new audio.
-    // This must happen BEFORE any early returns
+    // With the WebAudio backend, pause() accumulates playedDuration internally.
-    debug.log('Audio URL changed to:', audioUrl);
+    // seekTo(0) resets it so the new track starts from the beginning.
-
+    debug.log('Loading new audio URL:', audioUrl);
-    // COMPLETELY stop and destroy the current audio
     try {
-      // First pause if playing
       if (wavesurfer.isPlaying()) {
-        debug.log('Pausing current playback');
         wavesurfer.pause();
       }
-
+      wavesurfer.seekTo(0);
-      // Use empty() to completely destroy the waveform and reset media
+    } catch (err) {
-      debug.log('Calling wavesurfer.empty() to destroy audio');
+      debug.error('Error resetting before load:', err);
-      wavesurfer.empty();
-    } catch (error) {
-      debug.error('Error stopping previous audio:', error);
-      // Continue anyway to load new audio
     }
 
-    // Reset loading state to allow new load (cancel any pending loads)
-    loadingRef.current = false;
-
-    // Now start the new load
     loadingRef.current = true;
     setIsLoading(true);
     setError(null);
     setCurrentTime(0);
     setDuration(0);
 
-    // Load new audio
-    debug.log('Starting new audio load for:', audioUrl);
     wavesurfer
       .load(audioUrl)
       .then(() => {
-        debug.log('Audio load promise resolved');
+        debug.log('Audio loaded into WaveSurfer');
-        // Don't set loading to false here - wait for 'ready' event
+        loadingRef.current = false;
       })
-      .catch((error) => {
+      .catch((err) => {
-        debug.error('Failed to load audio:', error);
+        debug.error('Failed to load audio:', err);
-        debug.error('Audio URL:', audioUrl);
         loadingRef.current = false;
         setIsLoading(false);
-        setError(`Failed to load audio: ${error instanceof Error ? error.message : String(error)}`);
+        setError(`Failed to load audio: ${err instanceof Error ? err.message : String(err)}`);
       });
-  }, [audioUrl, setCurrentTime, setDuration]);
+  }, [audioUrl, wsReady, setCurrentTime, setDuration]);
 
   // Sync play/pause state (only when user clicks play/pause button, not auto-sync)
   // This effect is kept for external state changes but should be minimal
@@ -520,7 +318,6 @@ export function AudioPlayer() {
     if (!wavesurferRef.current || duration === 0) return;
 
     if (isPlaying && wavesurferRef.current.isPlaying() === false) {
-      // Only auto-play if audio is ready
       wavesurferRef.current.play().catch((error) => {
         debug.error('Failed to play:', error);
         setIsPlaying(false);
@@ -534,14 +331,7 @@ export function AudioPlayer() {
   // Sync volume
   useEffect(() => {
     if (wavesurferRef.current) {
-      // If using native playback, keep WaveSurfer muted regardless of volume setting
+      wavesurferRef.current.setVolume(volume);
-      if (isUsingNativePlaybackRef.current) {
-        wavesurferRef.current.setVolume(0);
-        debug.log('Volume sync: Using native playback, keeping WaveSurfer muted');
-      } else {
-        wavesurferRef.current.setVolume(volume);
-        debug.log('Volume synced:', volume);
-      }
     }
   }, [volume]);
 
@@ -566,7 +356,6 @@ export function AudioPlayer() {
       return;
     }
 
-    // Reset to beginning and play
     debug.log('Restarting current audio from beginning');
     wavesurfer.seekTo(0);
     wavesurfer.play().catch((error) => {
@@ -575,34 +364,35 @@ export function AudioPlayer() {
       setError(`Playback error: ${error instanceof Error ? error.message : String(error)}`);
     });
 
-    // Clear the restart flag
     clearRestartFlag();
   }, [shouldRestart, duration, setIsPlaying, clearRestartFlag]);
 
-  // Handle shouldAutoPlay flag - for story mode auto-advance
+  // Auto-play is handled exclusively in the WaveSurfer 'ready' event handler.
-  const shouldAutoPlay = usePlayerStore((state) => state.shouldAutoPlay);
+  // A separate effect here would race with the ready event since the WebAudio
-  const clearAutoPlayFlag = usePlayerStore((state) => state.clearAutoPlayFlag);
+  // backend needs to fully decode the audio before play() works correctly.
 
+  // Spacebar to play/pause (capture phase so it fires before focused elements)
   useEffect(() => {
-    const wavesurfer = wavesurferRef.current;
+    const onKeyDown = (e: KeyboardEvent) => {
-    if (!wavesurfer || !shouldAutoPlay || duration === 0) {
+      if (e.code !== 'Space') return;
-      return;
+      // Ignore if user is typing in an input/textarea
-    }
+      const tag = (e.target as HTMLElement)?.tagName;
-
+      if (tag === 'INPUT' || tag === 'TEXTAREA' || (e.target as HTMLElement)?.isContentEditable) {
-    // Auto-play the newly loaded audio
+        return;
-    debug.log('Auto-playing next track in story mode');
+      }
-    wavesurfer.seekTo(0);
+      if (audioUrl && duration > 0 && wavesurferRef.current) {
-    wavesurfer.play().catch((error) => {
+        e.preventDefault();
-      debug.error('Failed to auto-play:', error);
+        e.stopPropagation();
-      setIsPlaying(false);
+        if (wavesurferRef.current.isPlaying()) {
-      setError(`Playback error: ${error instanceof Error ? error.message : String(error)}`);
+          wavesurferRef.current.pause();
-    });
+        } else {
-
+          wavesurferRef.current.play().catch((err) => debug.error('Spacebar play failed:', err));
-    // Clear the auto-play flag
+        }
-    clearAutoPlayFlag();
+      }
-  }, [shouldAutoPlay, duration, setIsPlaying, clearAutoPlayFlag]);
+    };
-
+    document.addEventListener('keydown', onKeyDown, true);
-  // Handle loop - WaveSurfer handles this via the 'finish' event
+    return () => document.removeEventListener('keydown', onKeyDown, true);
+  }, [audioUrl, duration]);
 
   const handlePlayPause = async () => {
     // Standard WaveSurfer playback (works for both normal and native playback modes)
@@ -741,32 +531,32 @@ export function AudioPlayer() {
             size="icon"
             onClick={handlePlayPause}
             disabled={isLoading || duration === 0}
-            className="shrink-0"
+            className={`shrink-0 -mt-2 ${isPlaying ? 'bg-accent text-accent-foreground' : ''}`}
             title={duration === 0 && !isLoading ? 'Audio not loaded' : ''}
             aria-label={
               duration === 0 && !isLoading ? 'Audio not loaded' : isPlaying ? 'Pause' : 'Play'
             }
           >
-            {isPlaying ? <Pause className="h-5 w-5" /> : <Play className="h-5 w-5" />}
+            {isPlaying ? (
+              <Pause className="h-5 w-5 fill-current" />
+            ) : (
+              <Play className="h-5 w-5 fill-current" />
+            )}
           </Button>
 
           {/* Waveform */}
           <div className="flex-1 min-w-0 flex flex-col gap-1">
-            <div ref={waveformRef} className="w-full min-h-[80px]" />
+            <div ref={waveformRef} className="w-full min-h-[80px] select-none" />
-            {duration > 0 && (
+            <Slider
-              <Slider
+              value={duration > 0 ? [(currentTime / duration) * 100] : [0]}
-                value={duration > 0 ? [(currentTime / duration) * 100] : [0]}
+              onValueChange={handleSeek}
-                onValueChange={handleSeek}
+              max={100}
-                max={100}
+              step={0.1}
-                step={0.1}
+              className="w-full"
-                className="w-full"
+              aria-label="Playback position"
-                aria-label="Playback position"
+              aria-valuetext={`${formatAudioDuration(currentTime)} of ${formatAudioDuration(duration)}`}
-                aria-valuetext={`${formatAudioDuration(currentTime)} of ${formatAudioDuration(duration)}`}
+            />
-              />
+
-            )}
-            {isLoading && (
-              <div className="text-xs text-muted-foreground text-center py-2">Loading audio...</div>
-            )}
             {error && <div className="text-xs text-destructive text-center py-2">{error}</div>}
           </div>
 
@@ -777,19 +567,12 @@ export function AudioPlayer() {
             <span className="font-mono">{formatAudioDuration(duration)}</span>
           </div>
 
-          {/* Title */}
-          {title && (
-            <div className="text-sm font-medium truncate max-w-[200px] shrink-0 hidden lg:block">
-              {title}
-            </div>
-          )}
-
           {/* Loop Button */}
           <Button
             variant="ghost"
             size="icon"
             onClick={toggleLoop}
-            className={isLooping ? 'text-primary' : ''}
+            className={isLooping ? 'bg-accent text-accent-foreground' : ''}
             title="Toggle loop"
             aria-label={isLooping ? 'Stop looping' : 'Loop'}
           >

app/src/components/Generation/EngineModelSelector.tsx

@@ -20,6 +20,8 @@ const ENGINE_OPTIONS = [
   { value: 'luxtts', label: 'LuxTTS' },
   { value: 'chatterbox', label: 'Chatterbox' },
   { value: 'chatterbox_turbo', label: 'Chatterbox Turbo' },
+  { value: 'tada:1B', label: 'TADA 1B' },
+  { value: 'tada:3B', label: 'TADA 3B Multilingual' },
 ] as const;
 
 const ENGINE_DESCRIPTIONS: Record<string, string> = {
@@ -27,6 +29,7 @@ const ENGINE_DESCRIPTIONS: Record<string, string> = {
   luxtts: 'Fast, English-focused',
   chatterbox: '23 languages, incl. Hebrew',
   chatterbox_turbo: 'English, [laugh] [cough] tags',
+  tada: 'HumeAI, 700s+ coherent audio',
 };
 
 /** Engines that only support English and should force language to 'en' on select. */
@@ -34,6 +37,7 @@ const ENGLISH_ONLY_ENGINES = new Set(['luxtts', 'chatterbox_turbo']);
 
 function getSelectValue(engine: string, modelSize?: string): string {
   if (engine === 'qwen') return `qwen:${modelSize || '1.7B'}`;
+  if (engine === 'tada') return `tada:${modelSize || '1B'}`;
   return engine;
 }
 
@@ -48,6 +52,20 @@ function handleEngineChange(form: UseFormReturn<GenerationFormValues>, value: st
     if (!available.some((l) => l.value === currentLang)) {
       form.setValue('language', available[0]?.value ?? 'en');
     }
+  } else if (value.startsWith('tada:')) {
+    const [, modelSize] = value.split(':');
+    form.setValue('engine', 'tada');
+    form.setValue('modelSize', modelSize as '1B' | '3B');
+    // TADA 1B is English-only; 3B is multilingual
+    if (modelSize === '1B') {
+      form.setValue('language', 'en');
+    } else {
+      const currentLang = form.getValues('language');
+      const available = getLanguageOptionsForEngine('tada');
+      if (!available.some((l) => l.value === currentLang)) {
+        form.setValue('language', available[0]?.value ?? 'en');
+      }
+    }
   } else {
     form.setValue('engine', value as GenerationFormValues['engine']);
     form.setValue('modelSize', undefined as unknown as '1.7B' | '0.6B');

app/src/components/History/HistoryTable.tsx

@@ -15,7 +15,7 @@ import {
   Wand2,
 } from 'lucide-react';
 import { useEffect, useRef, useState } from 'react';
-import Loader from 'react-loaders';
+
 import { EffectsChainEditor } from '@/components/Effects/EffectsChainEditor';
 import { Button } from '@/components/ui/button';
 import {
@@ -56,8 +56,35 @@ import { formatDate, formatDuration, formatEngineName } from '@/lib/utils/format
 import { useGenerationStore } from '@/stores/generationStore';
 import { usePlayerStore } from '@/stores/playerStore';
 
-// OLD TABLE-BASED COMPONENT - REMOVED (can be found in git history)
+// ─── Audio Bars ─────────────────────────────────────────────────────────────
-// This is the new alternate history view with fixed height rows
+
+function AudioBars({ mode }: { mode: 'idle' | 'generating' | 'playing' }) {
+  const barColor = mode !== 'idle' ? 'bg-accent' : 'bg-muted-foreground/40';
+  return (
+    <div className="flex items-center gap-[2px] h-5">
+      {[0, 1, 2, 3, 4].map((i) => (
+        <motion.div
+          key={`${mode}-${i}`}
+          className={`w-[3px] rounded-full ${barColor}`}
+          animate={
+            mode === 'generating'
+              ? { height: ['6px', '16px', '6px'] }
+              : mode === 'playing'
+                ? { height: ['8px', '14px', '4px', '12px', '8px'] }
+                : { height: '8px' }
+          }
+          transition={
+            mode === 'generating'
+              ? { duration: 0.6, repeat: Infinity, delay: i * 0.08, ease: 'easeInOut' }
+              : mode === 'playing'
+                ? { duration: 1.2, repeat: Infinity, delay: i * 0.15, ease: 'easeInOut' }
+                : { duration: 0.4, ease: 'easeOut' }
+          }
+        />
+      ))}
+    </div>
+  );
+}
 
 // NEW ALTERNATE HISTORY VIEW - FIXED HEIGHT ROWS WITH INFINITE SCROLL
 export function HistoryTable() {
@@ -126,7 +153,9 @@ export function HistoryTable() {
     }
   }, [historyData, page]);
 
-  // Reset to page 0 when deletions or imports occur
+  // Reset to page 0 when deletions, imports, or generation completions occur
+  const pendingCount = useGenerationStore((state) => state.pendingGenerationIds.size);
+  const prevPendingCountRef = useRef(pendingCount);
   useEffect(() => {
     if (deleteGeneration.isSuccess || importGeneration.isSuccess) {
       setPage(0);
@@ -134,6 +163,19 @@ export function HistoryTable() {
     }
   }, [deleteGeneration.isSuccess, importGeneration.isSuccess]);
 
+  useEffect(() => {
+    // A generation finished (pending count decreased) — scroll back to show it
+    if (
+      prevPendingCountRef.current > 0 &&
+      pendingCount < prevPendingCountRef.current &&
+      page !== 0
+    ) {
+      setPage(0);
+      setAllHistory([]);
+    }
+    prevPendingCountRef.current = pendingCount;
+  }, [pendingCount, page]);
+
   // Intersection Observer for infinite scroll
   useEffect(() => {
     const loadMoreEl = loadMoreRef.current;
@@ -413,7 +455,7 @@ export function HistoryTable() {
                     role={isPlayable ? 'button' : undefined}
                     tabIndex={isPlayable ? 0 : undefined}
                     className={cn(
-                      'flex items-stretch gap-4 h-26 p-3',
+                      'flex items-stretch gap-4 h-26 p-3 outline-none',
                       isPlayable && 'hover:bg-muted/70 cursor-pointer rounded-md',
                       isVersionsExpanded && 'rounded-b-none',
                     )}
@@ -446,12 +488,9 @@ export function HistoryTable() {
                   >
                     {/* Status icon */}
                     <div className="flex items-center shrink-0 w-10 justify-center overflow-hidden">
-                      <div className="scale-50">
+                      <AudioBars
-                        <Loader
+                        mode={isGenerating ? 'generating' : isCurrentlyPlaying ? 'playing' : 'idle'}
-                          type={isGenerating ? 'line-scale' : 'line-scale-pulse-out-rapid'}
+                      />
-                          active={isGenerating || isCurrentlyPlaying}
-                        />
-                      </div>
                     </div>
 
                     {/* Left side - Meta information */}

app/src/components/ServerSettings/GpuAcceleration.tsx

@@ -243,7 +243,40 @@ export function GpuAcceleration() {
 
         {/* Native GPU detected - no CUDA download needed */}
 
-        {/* CUDA download section - only show when no GPU is active (native or CUDA) */}
+        {/* Currently running CUDA - show switch back to CPU */}
+        {isCurrentlyCuda && platform.metadata.isTauri && (
+          <>
+            {restartPhase !== 'idle' ? (
+              <div className="flex items-center gap-2 p-3 rounded-lg bg-primary/5 border">
+                <Loader2 className="h-4 w-4 animate-spin" />
+                <span className="text-sm">
+                  {restartPhase === 'stopping' && 'Stopping server...'}
+                  {restartPhase === 'waiting' && 'Restarting server...'}
+                  {restartPhase === 'ready' && 'Server restarted successfully!'}
+                </span>
+              </div>
+            ) : (
+              <div className="space-y-3">
+                <p className="text-sm text-muted-foreground">
+                  Running with CUDA GPU acceleration. Switch back to CPU if needed (you can
+                  re-download later).
+                </p>
+                <Button onClick={handleSwitchToCpu} variant="outline" className="w-full" size="sm">
+                  <RotateCw className="h-4 w-4 mr-2" />
+                  Switch to CPU Backend
+                </Button>
+              </div>
+            )}
+            {error && (
+              <div className="flex items-center gap-2 text-sm text-destructive">
+                <AlertCircle className="h-4 w-4 shrink-0" />
+                <span>{error}</span>
+              </div>
+            )}
+          </>
+        )}
+
+        {/* CUDA download/manage section - show when no native GPU and not currently running CUDA */}
         {!hasNativeGpu && !isCurrentlyCuda && (
           <>
             {/* Download progress (manual download or auto-update) */}
@@ -315,7 +348,7 @@ export function GpuAcceleration() {
                 )}
 
                 {/* Downloaded but not active - show switch button */}
-                {cudaAvailable && !isCurrentlyCuda && platform.metadata.isTauri && (
+                {cudaAvailable && platform.metadata.isTauri && (
                   <div className="space-y-3">
                     <p className="text-sm text-muted-foreground">
                       CUDA backend is downloaded and ready. Restart the server to enable GPU
@@ -328,27 +361,8 @@ export function GpuAcceleration() {
                   </div>
                 )}
 
-                {/* Currently active - show switch back to CPU */}
-                {isCurrentlyCuda && platform.metadata.isTauri && (
-                  <div className="space-y-3">
-                    <p className="text-sm text-muted-foreground">
-                      Running with CUDA GPU acceleration. Switch back to CPU if needed (you can
-                      re-download later).
-                    </p>
-                    <Button
-                      onClick={handleSwitchToCpu}
-                      variant="outline"
-                      className="w-full"
-                      size="sm"
-                    >
-                      <RotateCw className="h-4 w-4 mr-2" />
-                      Switch to CPU Backend
-                    </Button>
-                  </div>
-                )}
-
                 {/* Delete option when downloaded (and not active) */}
-                {cudaAvailable && !isCurrentlyCuda && (
+                {cudaAvailable && (
                   <Button
                     onClick={handleDelete}
                     variant="ghost"

app/src/components/ServerSettings/ModelManagement.tsx

@@ -62,6 +62,10 @@ const MODEL_DESCRIPTIONS: Record<string, string> = {
     'Production-grade open source TTS by Resemble AI. Supports 23 languages with voice cloning and emotion exaggeration control.',
   'chatterbox-turbo':
     'Streamlined 350M parameter TTS by Resemble AI. High-quality English speech with less compute and VRAM than larger models.',
+  'tada-1b':
+    'HumeAI TADA 1B — English speech-language model built on Llama 3.2 1B. Generates 700s+ of coherent audio with synchronized text-acoustic alignment.',
+  'tada-3b-ml':
+    'HumeAI TADA 3B Multilingual — built on Llama 3.2 3B. Supports 10 languages with high-fidelity voice cloning via text-acoustic dual alignment.',
   'whisper-base':
     'Smallest Whisper model (74M parameters). Fast transcription with moderate accuracy.',
   'whisper-small':
@@ -391,7 +395,8 @@ export function ModelManagement() {
       (m) =>
         m.model_name.startsWith('qwen-tts') ||
         m.model_name.startsWith('luxtts') ||
-        m.model_name.startsWith('chatterbox'),
+        m.model_name.startsWith('chatterbox') ||
+        m.model_name.startsWith('tada'),
     ) ?? [];
   const whisperModels = modelStatus?.models.filter((m) => m.model_name.startsWith('whisper')) ?? [];
 

app/src/components/ServerTab/AboutPage.tsx

@@ -0,0 +1,135 @@
+import { ArrowUpRight } from 'lucide-react';
+import type { CSSProperties, ReactNode } from 'react';
+import { useEffect, useState } from 'react';
+import voiceboxLogo from '@/assets/voicebox-logo.png';
+import { usePlatform } from '@/platform/PlatformContext';
+
+function FadeIn({ delay = 0, children }: { delay?: number; children: ReactNode }) {
+  return (
+    <div
+      className="animate-[fadeInUp_0.5s_ease_both]"
+      style={{ animationDelay: `${delay}ms` } as CSSProperties}
+    >
+      {children}
+    </div>
+  );
+}
+
+export function AboutPage() {
+  const platform = usePlatform();
+  const [version, setVersion] = useState('');
+
+  useEffect(() => {
+    platform.metadata
+      .getVersion()
+      .then(setVersion)
+      .catch(() => setVersion(''));
+  }, [platform]);
+
+  return (
+    <>
+      <style>{`
+        @keyframes fadeInUp {
+          from {
+            opacity: 0;
+            transform: translateY(8px);
+          }
+          to {
+            opacity: 1;
+            transform: translateY(0);
+          }
+        }
+      `}</style>
+      <div className="max-w-md mx-auto h-full flex items-center">
+        <div className="flex flex-col items-center text-center space-y-5">
+          <FadeIn delay={0}>
+            <img src={voiceboxLogo} alt="Voicebox" className="w-20 h-20 object-contain" />
+          </FadeIn>
+
+          <FadeIn delay={80}>
+            <div className="space-y-1.5">
+              <h1 className="text-lg font-semibold">Voicebox</h1>
+              <p className="text-xs text-muted-foreground/60 h-4">
+                {version ? `v${version}` : '\u00A0'}
+              </p>
+            </div>
+          </FadeIn>
+
+          <FadeIn delay={160}>
+            <p className="text-sm text-muted-foreground leading-relaxed max-w-sm">
+              The open-source voice synthesis studio. Clone voices, generate speech, apply effects,
+              and build voice-powered apps — all running locally on your machine.
+            </p>
+          </FadeIn>
+
+          <FadeIn delay={240}>
+            <div className="flex items-center gap-1.5 text-sm text-muted-foreground">
+              <span>Created by</span>
+              <a
+                href="https://github.com/jamiepine"
+                target="_blank"
+                rel="noopener noreferrer"
+                className="text-accent hover:underline"
+              >
+                Jamie Pine
+              </a>
+            </div>
+          </FadeIn>
+
+          <FadeIn delay={320}>
+            <div className="flex flex-wrap justify-center gap-3 pt-2">
+              <a
+                href="https://buymeacoffee.com/jamiepine"
+                target="_blank"
+                rel="noopener noreferrer"
+                className="group inline-flex items-center gap-2 rounded-lg border border-border/60 px-4 py-2 text-sm transition-colors hover:bg-muted/50"
+              >
+                <svg
+                  className="h-4 w-4 text-[#FFDD00]"
+                  viewBox="0 0 24 24"
+                  fill="currentColor"
+                  aria-hidden="true"
+                >
+                  <path d="m20.216 6.415-.132-.666c-.119-.598-.388-1.163-1.001-1.379-.197-.069-.42-.098-.57-.241-.152-.143-.196-.366-.231-.572-.065-.378-.125-.756-.192-1.133-.057-.325-.102-.69-.25-.987-.195-.4-.597-.634-.996-.788a5.723 5.723 0 0 0-.626-.194c-1-.263-2.05-.36-3.077-.416a25.834 25.834 0 0 0-3.7.062c-.915.083-1.88.184-2.75.5-.318.116-.646.256-.888.501-.297.302-.393.77-.177 1.146.154.267.415.456.692.58.36.162.737.284 1.123.366 1.075.238 2.189.331 3.287.37 1.218.05 2.437.01 3.65-.118.299-.033.598-.073.896-.119.352-.054.578-.513.474-.834-.124-.383-.457-.531-.834-.473-.466.074-.96.108-1.382.146-1.177.08-2.358.082-3.536.006a22.228 22.228 0 0 1-1.157-.107c-.086-.01-.18-.025-.258-.036-.243-.036-.484-.08-.724-.13-.111-.027-.111-.185 0-.212h.005c.277-.06.557-.108.838-.147h.002c.131-.009.263-.032.394-.048a25.076 25.076 0 0 1 3.426-.12c.674.019 1.347.067 2.017.144l.228.031c.267.04.533.088.798.145.392.085.895.113 1.07.542.055.137.08.288.111.431l.319 1.484a.237.237 0 0 1-.199.284h-.003c-.037.006-.075.01-.112.015a36.704 36.704 0 0 1-4.743.295 37.059 37.059 0 0 1-4.699-.304c-.14-.017-.293-.042-.417-.06-.326-.048-.649-.108-.973-.161-.393-.065-.768-.032-1.123.161-.29.16-.527.404-.675.701-.154.316-.199.66-.267 1-.069.34-.176.707-.135 1.056.087.753.613 1.365 1.37 1.502a39.69 39.69 0 0 0 11.343.376.483.483 0 0 1 .535.53l-.071.697-1.018 9.907c-.041.41-.047.832-.125 1.237-.122.637-.553 1.028-1.182 1.171-.577.131-1.165.2-1.756.205-.656.004-1.31-.025-1.966-.022-.699.004-1.556-.06-2.095-.58-.475-.458-.54-1.174-.605-1.793l-.731-7.013-.322-3.094c-.037-.351-.286-.695-.678-.678-.336.015-.718.3-.678.679l.228 2.185.949 9.112c.147 1.344 1.174 2.068 2.446 2.272.742.12 1.503.144 2.257.156.966.016 1.942.053 2.892-.122 1.408-.258 2.465-1.198 2.616-2.657.34-3.332.683-6.663 1.024-9.995l.215-2.087a.484.484 0 0 1 .39-.426c.402-.078.787-.212 1.074-.518.455-.488.546-1.124.385-1.766zm-1.478.772c-.145.137-.363.201-.578.233-2.416.359-4.866.54-7.308.46-1.748-.06-3.477-.254-5.207-.498-.17-.024-.353-.055-.47-.18-.22-.236-.111-.71-.054-.995.052-.26.152-.609.463-.646.484-.057 1.046.148 1.526.22.577.088 1.156.159 1.737.212 2.48.226 5.002.19 7.472-.14.45-.06.899-.13 1.345-.21.399-.072.84-.206 1.08.206.166.281.188.657.162.974a.544.544 0 0 1-.169.364zm-6.159 3.9c-.862.37-1.84.788-3.109.788a5.884 5.884 0 0 1-1.569-.217l.877 9.004c.065.78.717 1.38 1.5 1.38 0 0 1.243.065 1.658.065.447 0 1.786-.065 1.786-.065.783 0 1.434-.6 1.499-1.38l.94-9.95a3.996 3.996 0 0 0-1.322-.238c-.826 0-1.491.284-2.26.613z" />
+                </svg>
+                Buy me a coffee
+                <ArrowUpRight className="h-3.5 w-3.5 text-muted-foreground/40 group-hover:text-muted-foreground transition-colors" />
+              </a>
+              <a
+                href="https://github.com/jamiepine/voicebox"
+                target="_blank"
+                rel="noopener noreferrer"
+                className="group inline-flex items-center gap-2 rounded-lg border border-border/60 px-4 py-2 text-sm transition-colors hover:bg-muted/50"
+              >
+                <svg
+                  className="h-4 w-4 text-muted-foreground"
+                  viewBox="0 0 24 24"
+                  fill="currentColor"
+                  aria-hidden="true"
+                >
+                  <path d="M12 2C6.477 2 2 6.484 2 12.017c0 4.425 2.865 8.18 6.839 9.504.5.092.682-.217.682-.483 0-.237-.008-.868-.013-1.703-2.782.605-3.369-1.343-3.369-1.343-.454-1.158-1.11-1.466-1.11-1.466-.908-.62.069-.608.069-.608 1.003.07 1.531 1.032 1.531 1.032.892 1.53 2.341 1.088 2.91.832.092-.647.35-1.088.636-1.338-2.22-.253-4.555-1.113-4.555-4.951 0-1.093.39-1.988 1.029-2.688-.103-.253-.446-1.272.098-2.65 0 0 .84-.27 2.75 1.026A9.564 9.564 0 0112 6.844c.85.004 1.705.115 2.504.337 1.909-1.296 2.747-1.027 2.747-1.027.546 1.379.202 2.398.1 2.651.64.7 1.028 1.595 1.028 2.688 0 3.848-2.339 4.695-4.566 4.943.359.309.678.92.678 1.855 0 1.338-.012 2.419-.012 2.747 0 .268.18.58.688.482A10.019 10.019 0 0022 12.017C22 6.484 17.522 2 12 2z" />
+                </svg>
+                GitHub
+                <ArrowUpRight className="h-3.5 w-3.5 text-muted-foreground/40 group-hover:text-muted-foreground transition-colors" />
+              </a>
+            </div>
+          </FadeIn>
+
+          <FadeIn delay={400}>
+            <p className="text-xs text-muted-foreground/40 pt-4">
+              Licensed under{' '}
+              <a
+                href="https://github.com/jamiepine/voicebox/blob/main/LICENSE"
+                target="_blank"
+                rel="noopener noreferrer"
+                className="hover:text-muted-foreground/60 transition-colors"
+              >
+                MIT
+              </a>
+            </p>
+          </FadeIn>
+        </div>
+      </div>
+    </>
+  );
+}

app/src/components/ServerTab/ChangelogPage.tsx

@@ -0,0 +1,220 @@
+import changelogRaw from 'virtual:changelog';
+import { useMemo, useState } from 'react';
+import { Badge } from '@/components/ui/badge';
+import { type ChangelogEntry, parseChangelog } from '@/lib/utils/parseChangelog';
+
+function renderMarkdown(md: string): React.ReactNode[] {
+  const lines = md.split('\n');
+  const elements: React.ReactNode[] = [];
+  let i = 0;
+
+  while (i < lines.length) {
+    const line = lines[i];
+
+    // Skip empty lines
+    if (line.trim() === '') {
+      i++;
+      continue;
+    }
+
+    // Tables — collect all lines starting with |
+    if (line.trim().startsWith('|')) {
+      const tableLines: string[] = [];
+      while (i < lines.length && lines[i].trim().startsWith('|')) {
+        tableLines.push(lines[i]);
+        i++;
+      }
+      elements.push(renderTable(tableLines, elements.length));
+      continue;
+    }
+
+    // Headings
+    if (line.startsWith('#### ')) {
+      elements.push(
+        <h5 key={elements.length} className="text-sm font-medium mt-5 mb-1">
+          {inlineMarkdown(line.slice(5))}
+        </h5>,
+      );
+      i++;
+      continue;
+    }
+    if (line.startsWith('### ')) {
+      elements.push(
+        <h4 key={elements.length} className="text-sm font-medium mt-6 mb-2">
+          {inlineMarkdown(line.slice(4))}
+        </h4>,
+      );
+      i++;
+      continue;
+    }
+
+    // List items — collect consecutive
+    if (line.startsWith('- ')) {
+      const items: string[] = [];
+      while (i < lines.length && lines[i].startsWith('- ')) {
+        items.push(lines[i].slice(2));
+        i++;
+      }
+      elements.push(
+        <ul key={elements.length} className="space-y-1 my-2">
+          {items.map((item, idx) => (
+            <li key={idx} className="text-sm text-muted-foreground flex gap-2">
+              <span className="text-muted-foreground/50 select-none shrink-0">&bull;</span>
+              <span>{inlineMarkdown(item)}</span>
+            </li>
+          ))}
+        </ul>,
+      );
+      continue;
+    }
+
+    // Paragraph
+    elements.push(
+      <p key={elements.length} className="text-sm text-muted-foreground my-2">
+        {inlineMarkdown(line)}
+      </p>,
+    );
+    i++;
+  }
+
+  return elements;
+}
+
+function renderTable(tableLines: string[], keyBase: number): React.ReactNode {
+  const parseRow = (line: string) =>
+    line
+      .split('|')
+      .slice(1, -1)
+      .map((c) => c.trim());
+
+  const headers = parseRow(tableLines[0]);
+  // Skip separator line (index 1)
+  const rows = tableLines.slice(2).map(parseRow);
+
+  return (
+    <div key={keyBase} className="overflow-x-auto my-3">
+      <table className="text-sm w-full">
+        <thead>
+          <tr className="border-b">
+            {headers.map((h, hIdx) => (
+              <th
+                key={hIdx}
+                className="text-left py-1.5 pr-4 text-muted-foreground font-medium text-xs"
+              >
+                {inlineMarkdown(h)}
+              </th>
+            ))}
+          </tr>
+        </thead>
+        <tbody>
+          {rows.map((row, rowIdx) => (
+            <tr key={rowIdx} className="border-b border-border/50">
+              {row.map((cell, cellIdx) => (
+                <td key={cellIdx} className="py-1.5 pr-4 text-muted-foreground">
+                  {inlineMarkdown(cell)}
+                </td>
+              ))}
+            </tr>
+          ))}
+        </tbody>
+      </table>
+    </div>
+  );
+}
+
+function inlineMarkdown(text: string): React.ReactNode {
+  // Process inline markdown: bold, code, links
+  const parts: React.ReactNode[] = [];
+  // Regex matches: **bold**, `code`, [text](url)
+  const inlineRe = /\*\*(.+?)\*\*|`([^`]+)`|\[([^\]]+)\]\(([^)]+)\)/g;
+  let lastIndex = 0;
+  let match: RegExpExecArray | null = inlineRe.exec(text);
+
+  while (match !== null) {
+    if (match.index > lastIndex) {
+      parts.push(text.slice(lastIndex, match.index));
+    }
+
+    if (match[1] !== undefined) {
+      // Bold
+      parts.push(
+        <strong key={parts.length} className="font-medium text-foreground">
+          {match[1]}
+        </strong>,
+      );
+    } else if (match[2] !== undefined) {
+      // Code
+      parts.push(
+        <code key={parts.length} className="px-1 py-0.5 rounded bg-muted text-xs font-mono">
+          {match[2]}
+        </code>,
+      );
+    } else if (match[3] !== undefined && match[4] !== undefined) {
+      // Link
+      parts.push(
+        <a
+          key={parts.length}
+          href={match[4]}
+          target="_blank"
+          rel="noopener noreferrer"
+          className="text-accent hover:underline"
+        >
+          {match[3]}
+        </a>,
+      );
+    }
+
+    lastIndex = match.index + match[0].length;
+    match = inlineRe.exec(text);
+  }
+
+  if (lastIndex < text.length) {
+    parts.push(text.slice(lastIndex));
+  }
+
+  return parts.length === 1 ? parts[0] : parts;
+}
+
+function ChangelogEntryCard({ entry }: { entry: ChangelogEntry }) {
+  const [expanded, setExpanded] = useState(false);
+  const content = useMemo(() => renderMarkdown(entry.body), [entry.body]);
+  const isLong = entry.body.split('\n').length > 12;
+
+  return (
+    <div className="border-b border-border/50 pb-6">
+      <div className="flex items-baseline gap-3 mb-1">
+        <h3 className="text-sm font-medium">{entry.version}</h3>
+        {entry.date && <span className="text-xs text-muted-foreground">{entry.date}</span>}
+        {entry.version === 'Unreleased' && <Badge variant="outline">dev</Badge>}
+      </div>
+
+      <div className={isLong && !expanded ? 'max-h-48 overflow-hidden relative' : ''}>
+        {content}
+        {isLong && !expanded && (
+          <div className="absolute bottom-0 left-0 right-0 h-16 bg-gradient-to-t from-background to-transparent" />
+        )}
+      </div>
+
+      {isLong && (
+        <button
+          onClick={() => setExpanded(!expanded)}
+          className="text-xs text-accent hover:underline mt-2"
+        >
+          {expanded ? 'Show less' : 'Show more'}
+        </button>
+      )}
+    </div>
+  );
+}
+
+export function ChangelogPage() {
+  const entries = useMemo(() => parseChangelog(changelogRaw), []);
+
+  return (
+    <div className="space-y-6 max-w-2xl">
+      {entries.map((entry) => (
+        <ChangelogEntryCard key={entry.version} entry={entry} />
+      ))}
+    </div>
+  );
+}

app/src/components/ServerTab/GeneralPage.tsx

@@ -0,0 +1,379 @@
+import { zodResolver } from '@hookform/resolvers/zod';
+import { AlertCircle, ArrowUpRight, Book, Download, Loader2, RefreshCw } from 'lucide-react';
+import { useEffect, useState } from 'react';
+import { useForm } from 'react-hook-form';
+import * as z from 'zod';
+import { Button } from '@/components/ui/button';
+import { Form, FormControl, FormField, FormItem, FormMessage } from '@/components/ui/form';
+import { Input } from '@/components/ui/input';
+import { Progress } from '@/components/ui/progress';
+import { Toggle } from '@/components/ui/toggle';
+import { useToast } from '@/components/ui/use-toast';
+import { useAutoUpdater } from '@/hooks/useAutoUpdater';
+import { useServerHealth } from '@/lib/hooks/useServer';
+import { usePlatform } from '@/platform/PlatformContext';
+import { useServerStore } from '@/stores/serverStore';
+import { SettingRow, SettingSection } from './SettingRow';
+
+const connectionSchema = z.object({
+  serverUrl: z.string().url('Please enter a valid URL'),
+});
+
+type ConnectionFormValues = z.infer<typeof connectionSchema>;
+
+export function GeneralPage() {
+  const platform = usePlatform();
+  const serverUrl = useServerStore((state) => state.serverUrl);
+  const setServerUrl = useServerStore((state) => state.setServerUrl);
+  const keepServerRunningOnClose = useServerStore((state) => state.keepServerRunningOnClose);
+  const setKeepServerRunningOnClose = useServerStore((state) => state.setKeepServerRunningOnClose);
+  const mode = useServerStore((state) => state.mode);
+  const setMode = useServerStore((state) => state.setMode);
+  const { toast } = useToast();
+  const { data: health, isLoading, error: healthError } = useServerHealth();
+
+  const form = useForm<ConnectionFormValues>({
+    resolver: zodResolver(connectionSchema),
+    defaultValues: { serverUrl },
+  });
+
+  useEffect(() => {
+    form.reset({ serverUrl });
+  }, [serverUrl, form]);
+
+  const { isDirty } = form.formState;
+
+  function onSubmit(data: ConnectionFormValues) {
+    setServerUrl(data.serverUrl);
+    form.reset(data);
+    toast({
+      title: 'Server URL updated',
+      description: `Connected to ${data.serverUrl}`,
+    });
+  }
+
+  return (
+    <div className="space-y-8 max-w-2xl">
+      <div className="grid grid-cols-2 gap-3">
+        <a
+          href="https://docs.voicebox.sh"
+          target="_blank"
+          rel="noopener noreferrer"
+          className="group flex items-center gap-3 rounded-lg border border-border/60 p-4 transition-colors hover:bg-muted/50"
+        >
+          <Book className="h-5 w-5 shrink-0 text-accent" strokeWidth={2.5} />
+          <div className="min-w-0 flex-1">
+            <div className="text-sm font-medium">Read the Docs</div>
+            <div className="text-xs text-muted-foreground">docs.voicebox.sh</div>
+          </div>
+          <ArrowUpRight className="h-4 w-4 text-muted-foreground/40 group-hover:text-muted-foreground transition-colors" />
+        </a>
+        <a
+          href="https://discord.gg/StkzQasqPS"
+          target="_blank"
+          rel="noopener noreferrer"
+          className="group flex items-center gap-3 rounded-lg border border-border/60 p-4 transition-colors hover:bg-muted/50"
+        >
+          <svg
+            className="h-5 w-5 shrink-0 text-accent"
+            viewBox="0 0 24 24"
+            fill="currentColor"
+            aria-hidden="true"
+          >
+            <path d="M20.317 4.37a19.791 19.791 0 0 0-4.885-1.515.074.074 0 0 0-.079.037c-.21.375-.444.864-.608 1.25a18.27 18.27 0 0 0-5.487 0 12.64 12.64 0 0 0-.617-1.25.077.077 0 0 0-.079-.037A19.736 19.736 0 0 0 3.677 4.37a.07.07 0 0 0-.032.027C.533 9.046-.32 13.58.099 18.057a.082.082 0 0 0 .031.057 19.9 19.9 0 0 0 5.993 3.03.078.078 0 0 0 .084-.028c.462-.63.874-1.295 1.226-1.994a.076.076 0 0 0-.041-.106 13.107 13.107 0 0 1-1.872-.892.077.077 0 0 1-.008-.128 10.2 10.2 0 0 0 .372-.292.074.074 0 0 1 .077-.01c3.928 1.793 8.18 1.793 12.062 0a.074.074 0 0 1 .078.01c.12.098.246.198.373.292a.077.077 0 0 1-.006.127 12.299 12.299 0 0 1-1.873.892.077.077 0 0 0-.041.107c.36.698.772 1.362 1.225 1.993a.076.076 0 0 0 .084.028 19.839 19.839 0 0 0 6.002-3.03.077.077 0 0 0 .032-.054c.5-5.177-.838-9.674-3.549-13.66a.061.061 0 0 0-.031-.03zM8.02 15.33c-1.183 0-2.157-1.085-2.157-2.419 0-1.333.956-2.419 2.157-2.419 1.21 0 2.176 1.095 2.157 2.42 0 1.333-.956 2.418-2.157 2.418zm7.975 0c-1.183 0-2.157-1.085-2.157-2.419 0-1.333.955-2.419 2.157-2.419 1.21 0 2.176 1.095 2.157 2.42 0 1.333-.946 2.418-2.157 2.418z" />
+          </svg>
+          <div className="min-w-0 flex-1">
+            <div className="text-sm font-medium">Join the Discord</div>
+            <div className="text-xs text-muted-foreground">Get help & share voices</div>
+          </div>
+          <ArrowUpRight className="h-4 w-4 text-muted-foreground/40 group-hover:text-muted-foreground transition-colors" />
+        </a>
+      </div>
+
+      <SettingSection>
+        <SettingRow
+          title="Server URL"
+          description="The address of your voicebox backend server."
+          action={
+            <ConnectionStatus health={health} isLoading={isLoading} healthError={healthError} />
+          }
+        >
+          <Form {...form}>
+            <form onSubmit={form.handleSubmit(onSubmit)} className="flex gap-2">
+              <FormField
+                control={form.control}
+                name="serverUrl"
+                render={({ field }) => (
+                  <FormItem className="flex-1">
+                    <FormControl>
+                      <Input placeholder="http://127.0.0.1:17493" {...field} />
+                    </FormControl>
+                    <FormMessage />
+                  </FormItem>
+                )}
+              />
+              {isDirty && (
+                <Button type="submit" size="sm">
+                  Save
+                </Button>
+              )}
+            </form>
+          </Form>
+        </SettingRow>
+
+        <SettingRow
+          title="Keep server running when app closes"
+          description="The server will continue running in the background after closing the app."
+          htmlFor="keepServerRunning"
+          action={
+            <Toggle
+              id="keepServerRunning"
+              checked={keepServerRunningOnClose}
+              onCheckedChange={(checked: boolean) => {
+                setKeepServerRunningOnClose(checked);
+                platform.lifecycle.setKeepServerRunning(checked).catch((error) => {
+                  console.error('Failed to sync setting to Rust:', error);
+                  setKeepServerRunningOnClose(!checked);
+                  toast({
+                    title: 'Failed to update setting',
+                    description: 'Could not sync setting to backend.',
+                    variant: 'destructive',
+                  });
+                  return;
+                });
+                toast({
+                  title: 'Setting updated',
+                  description: checked
+                    ? 'Server will continue running when app closes'
+                    : 'Server will stop when app closes',
+                });
+              }}
+            />
+          }
+        />
+
+        {platform.metadata.isTauri && (
+          <SettingRow
+            title="Allow network access"
+            description="Makes the server accessible from other devices on your network. Restart the app after changing."
+            htmlFor="allowNetworkAccess"
+            action={
+              <Toggle
+                id="allowNetworkAccess"
+                checked={mode === 'remote'}
+                onCheckedChange={(checked: boolean) => {
+                  setMode(checked ? 'remote' : 'local');
+                  toast({
+                    title: 'Setting updated',
+                    description: checked
+                      ? 'Network access enabled. Restart the app to apply.'
+                      : 'Network access disabled. Restart the app to apply.',
+                  });
+                }}
+              />
+            }
+          />
+        )}
+      </SettingSection>
+
+      <ApiReferenceCard serverUrl={serverUrl} />
+
+      {platform.metadata.isTauri && <UpdatesSection />}
+    </div>
+  );
+}
+
+function ConnectionStatus({
+  health,
+  isLoading,
+  healthError,
+}: {
+  health: ReturnType<typeof useServerHealth>['data'];
+  isLoading: boolean;
+  healthError: ReturnType<typeof useServerHealth>['error'];
+}) {
+  if (isLoading) {
+    return (
+      <div className="flex items-center gap-2 rounded-full border border-border/60 px-3 py-1">
+        <Loader2 className="h-3 w-3 animate-spin text-muted-foreground" />
+        <span className="text-xs text-muted-foreground">Connecting</span>
+      </div>
+    );
+  }
+  if (healthError) {
+    return (
+      <div className="flex items-center gap-2 rounded-full border border-destructive/30 px-3 py-1">
+        <span className="relative flex h-2 w-2">
+          <span className="absolute inline-flex h-full w-full rounded-full bg-destructive/40" />
+          <span className="relative inline-flex h-2 w-2 rounded-full bg-destructive" />
+        </span>
+        <span className="text-xs text-destructive">Offline</span>
+      </div>
+    );
+  }
+  if (health) {
+    return (
+      <div className="flex items-center gap-2 rounded-full border border-accent/30 px-3 py-1">
+        <span className="relative flex h-2 w-2">
+          <span className="absolute inline-flex h-full w-full animate-ping rounded-full bg-accent/60" />
+          <span className="relative inline-flex h-2 w-2 rounded-full bg-accent shadow-[0_0_6px_1px_hsl(var(--accent)/0.5)]" />
+        </span>
+        <span className="text-xs text-muted-foreground">Online</span>
+      </div>
+    );
+  }
+  return null;
+}
+
+function UpdatesSection() {
+  const platform = usePlatform();
+  const { status, checkForUpdates, downloadAndInstall, restartAndInstall } = useAutoUpdater(false);
+  const [currentVersion, setCurrentVersion] = useState<string>('');
+  const isDev = !import.meta.env?.PROD;
+
+  useEffect(() => {
+    platform.metadata
+      .getVersion()
+      .then(setCurrentVersion)
+      .catch(() => setCurrentVersion('Unknown'));
+  }, [platform]);
+
+  return (
+    <SettingSection title="App Updates" description={`v${currentVersion}${isDev ? ' (dev)' : ''}`}>
+      {isDev ? (
+        <SettingRow
+          title="Development mode"
+          description="Auto-updates are disabled in development mode."
+        />
+      ) : (
+        <>
+          <SettingRow
+            title="Check for updates"
+            description={
+              status.available
+                ? `Version ${status.version} available`
+                : status.checking
+                  ? 'Checking...'
+                  : "You're up to date"
+            }
+            action={
+              <Button
+                onClick={checkForUpdates}
+                disabled={status.checking || status.downloading || status.readyToInstall}
+                variant="outline"
+                size="sm"
+              >
+                <RefreshCw
+                  className={`h-3.5 w-3.5 mr-1.5 ${status.checking ? 'animate-spin' : ''}`}
+                />
+                Check
+              </Button>
+            }
+          />
+
+          {status.error && (
+            <SettingRow title="Update error">
+              <div className="flex items-center gap-2 text-sm text-destructive">
+                <AlertCircle className="h-4 w-4" />
+                {status.error}
+              </div>
+            </SettingRow>
+          )}
+
+          {status.available && !status.downloading && !status.readyToInstall && (
+            <SettingRow
+              title={`Update to ${status.version}`}
+              description="Download and install the latest version."
+              action={
+                <Button onClick={downloadAndInstall} size="sm">
+                  <Download className="h-3.5 w-3.5 mr-1.5" />
+                  Download
+                </Button>
+              }
+            />
+          )}
+
+          {status.downloading && (
+            <SettingRow title="Downloading update...">
+              <div className="space-y-1.5">
+                <Progress value={status.downloadProgress} />
+                <div className="flex items-center justify-between text-xs text-muted-foreground">
+                  {status.downloadedBytes !== undefined &&
+                  status.totalBytes !== undefined &&
+                  status.totalBytes > 0 ? (
+                    <span>
+                      {(status.downloadedBytes / 1024 / 1024).toFixed(1)} MB /{' '}
+                      {(status.totalBytes / 1024 / 1024).toFixed(1)} MB
+                    </span>
+                  ) : (
+                    <span />
+                  )}
+                  {status.downloadProgress !== undefined && <span>{status.downloadProgress}%</span>}
+                </div>
+              </div>
+            </SettingRow>
+          )}
+
+          {status.readyToInstall && (
+            <SettingRow
+              title="Update ready to install"
+              description={`Version ${status.version} has been downloaded. Restart to complete.`}
+              action={
+                <Button onClick={restartAndInstall} size="sm">
+                  <RefreshCw className="h-3.5 w-3.5 mr-1.5" />
+                  Restart Now
+                </Button>
+              }
+            />
+          )}
+        </>
+      )}
+    </SettingSection>
+  );
+}
+
+const API_ENDPOINTS = [
+  { method: 'POST', path: '/generate', label: 'Generate speech' },
+  { method: 'GET', path: '/health', label: 'Server status' },
+  { method: 'GET', path: '/profiles', label: 'List voices' },
+  { method: 'GET', path: '/history', label: 'Past generations' },
+];
+
+function ApiReferenceCard({ serverUrl }: { serverUrl: string }) {
+  return (
+    <div className="rounded-lg border border-border/60 p-4 space-y-3">
+      <div>
+        <h3 className="text-sm font-medium">API Access</h3>
+        <p className="text-sm text-muted-foreground">
+          Integrate Voicebox into your workflow via the REST API at{' '}
+          <code className="text-xs bg-muted px-1 py-0.5 rounded font-mono">{serverUrl}</code>
+        </p>
+      </div>
+      <div className="space-y-1">
+        {API_ENDPOINTS.map((ep) => (
+          <div key={ep.path} className="flex items-center gap-2.5 py-1">
+            <span
+              className={`text-[10px] font-mono font-semibold w-9 text-center rounded px-1 py-px ${
+                ep.method === 'POST' ? 'bg-accent/10 text-accent' : 'bg-muted text-muted-foreground'
+              }`}
+            >
+              {ep.method}
+            </span>
+            <code className="text-xs font-mono text-muted-foreground">{ep.path}</code>
+            <span className="text-xs text-muted-foreground/50 ml-auto">{ep.label}</span>
+          </div>
+        ))}
+      </div>
+      <p className="text-xs text-muted-foreground">
+        <a
+          href={`${serverUrl}/docs`}
+          target="_blank"
+          rel="noopener noreferrer"
+          className="text-accent hover:underline"
+        >
+          View the full API reference
+        </a>
+      </p>
+    </div>
+  );
+}

app/src/components/ServerTab/GenerationPage.tsx

@@ -0,0 +1,138 @@
+import { FolderOpen } from 'lucide-react';
+import { useCallback, useEffect, useState } from 'react';
+import { Button } from '@/components/ui/button';
+import { Slider } from '@/components/ui/slider';
+import { Toggle } from '@/components/ui/toggle';
+import { usePlatform } from '@/platform/PlatformContext';
+import { useServerStore } from '@/stores/serverStore';
+import { SettingRow, SettingSection } from './SettingRow';
+
+export function GenerationPage() {
+  const platform = usePlatform();
+  const serverUrl = useServerStore((state) => state.serverUrl);
+  const maxChunkChars = useServerStore((state) => state.maxChunkChars);
+  const setMaxChunkChars = useServerStore((state) => state.setMaxChunkChars);
+  const crossfadeMs = useServerStore((state) => state.crossfadeMs);
+  const setCrossfadeMs = useServerStore((state) => state.setCrossfadeMs);
+  const normalizeAudio = useServerStore((state) => state.normalizeAudio);
+  const setNormalizeAudio = useServerStore((state) => state.setNormalizeAudio);
+  const autoplayOnGenerate = useServerStore((state) => state.autoplayOnGenerate);
+  const setAutoplayOnGenerate = useServerStore((state) => state.setAutoplayOnGenerate);
+  const [opening, setOpening] = useState(false);
+  const [generationsPath, setGenerationsPath] = useState<string | null>(null);
+
+  useEffect(() => {
+    fetch(`${serverUrl}/health/filesystem`)
+      .then((res) => res.json())
+      .then((data) => {
+        const genDir = data.directories?.find((d: { path: string }) =>
+          d.path.includes('generations'),
+        );
+        if (genDir?.path) setGenerationsPath(genDir.path);
+      })
+      .catch(() => {});
+  }, [serverUrl]);
+
+  const openGenerationsFolder = useCallback(async () => {
+    if (!generationsPath) return;
+    setOpening(true);
+    try {
+      await platform.filesystem.openPath(generationsPath);
+    } catch (e) {
+      console.error('Failed to open generations folder:', e);
+    } finally {
+      setOpening(false);
+    }
+  }, [platform, generationsPath]);
+
+  return (
+    <div className="space-y-8 max-w-2xl">
+      <SettingSection
+        title="Generation"
+        description="Controls for long text generation. These settings apply to all engines."
+      >
+        <SettingRow
+          title="Auto-chunking limit"
+          description="Long text is split into chunks at sentence boundaries. Lower values can improve quality for long outputs."
+          action={
+            <span className="text-sm tabular-nums text-muted-foreground">
+              {maxChunkChars} chars
+            </span>
+          }
+        >
+          <Slider
+            id="maxChunkChars"
+            value={[maxChunkChars]}
+            onValueChange={([value]) => setMaxChunkChars(value)}
+            min={100}
+            max={5000}
+            step={50}
+            aria-label="Auto-chunking character limit"
+          />
+        </SettingRow>
+
+        <SettingRow
+          title="Chunk crossfade"
+          description="Blends audio between chunks to smooth transitions. Set to 0 for a hard cut."
+          action={
+            <span className="text-sm tabular-nums text-muted-foreground">
+              {crossfadeMs === 0 ? 'Cut' : `${crossfadeMs}ms`}
+            </span>
+          }
+        >
+          <Slider
+            id="crossfadeMs"
+            value={[crossfadeMs]}
+            onValueChange={([value]) => setCrossfadeMs(value)}
+            min={0}
+            max={200}
+            step={10}
+            aria-label="Chunk crossfade duration"
+          />
+        </SettingRow>
+
+        <SettingRow
+          title="Normalize audio"
+          description="Adjusts output volume to a consistent level across generations."
+          htmlFor="normalizeAudio"
+          action={
+            <Toggle
+              id="normalizeAudio"
+              checked={normalizeAudio}
+              onCheckedChange={setNormalizeAudio}
+            />
+          }
+        />
+
+        <SettingRow
+          title="Autoplay on generate"
+          description="Automatically play audio when a generation completes."
+          htmlFor="autoplayOnGenerate"
+          action={
+            <Toggle
+              id="autoplayOnGenerate"
+              checked={autoplayOnGenerate}
+              onCheckedChange={setAutoplayOnGenerate}
+            />
+          }
+        />
+
+        <SettingRow
+          title="Generations folder"
+          description={generationsPath ?? 'Where generated audio files are stored on disk.'}
+          action={
+            <Button
+              variant="outline"
+              size="sm"
+              onClick={openGenerationsFolder}
+              disabled={opening || !generationsPath}
+            >
+              <FolderOpen className="h-3.5 w-3.5 mr-1.5" />
+              Open
+            </Button>
+          }
+        />
+      </SettingSection>
+    </div>
+  );
+}

app/src/components/ServerTab/GpuPage.tsx

@@ -0,0 +1,405 @@
+import { useQuery, useQueryClient } from '@tanstack/react-query';
+import { AlertCircle, Cpu, Download, Loader2, RotateCw, Trash2 } from 'lucide-react';
+import { useCallback, useEffect, useRef, useState } from 'react';
+import { Button } from '@/components/ui/button';
+import { Progress } from '@/components/ui/progress';
+import { apiClient } from '@/lib/api/client';
+import type { CudaDownloadProgress, HealthResponse } from '@/lib/api/types';
+import { useServerHealth } from '@/lib/hooks/useServer';
+import { usePlatform } from '@/platform/PlatformContext';
+import { useServerStore } from '@/stores/serverStore';
+import { SettingRow, SettingSection } from './SettingRow';
+
+type RestartPhase = 'idle' | 'stopping' | 'waiting' | 'ready';
+
+function AppleLogo({ className }: { className?: string }) {
+  return (
+    <svg className={className} viewBox="0 0 24 24" fill="currentColor" aria-hidden="true">
+      <path d="M18.71 19.5c-.83 1.24-1.71 2.45-3.05 2.47-1.34.03-1.77-.79-3.29-.79-1.53 0-2 .77-3.27.82-1.31.05-2.3-1.32-3.14-2.53C4.25 17 2.94 12.45 4.7 9.39c.87-1.52 2.43-2.48 4.12-2.51 1.28-.02 2.5.87 3.29.87.78 0 2.26-1.07 3.8-.91.65.03 2.47.26 3.64 1.98-.09.06-2.17 1.28-2.15 3.81.03 3.02 2.65 4.03 2.68 4.04-.03.07-.42 1.44-1.38 2.83M13 3.5c.73-.83 1.94-1.46 2.94-1.5.13 1.17-.34 2.35-1.04 3.19-.69.85-1.83 1.51-2.95 1.42-.15-1.15.41-2.35 1.05-3.11z" />
+    </svg>
+  );
+}
+
+function GpuIcon({ className }: { className?: string }) {
+  return (
+    <svg
+      className={className}
+      viewBox="0 0 24 24"
+      fill="none"
+      stroke="currentColor"
+      strokeWidth="1.5"
+      strokeLinecap="round"
+      strokeLinejoin="round"
+      aria-hidden="true"
+    >
+      <rect x="4" y="6" width="16" height="12" rx="2" />
+      <path d="M2 10h2M2 14h2M20 10h2M20 14h2" />
+      <path d="M9 10h6M9 14h4" />
+    </svg>
+  );
+}
+
+function GpuInfoCard({ health }: { health: HealthResponse }) {
+  const hasGpu = health.gpu_available && health.gpu_type;
+
+  // Parse GPU name from type string like "CUDA (NVIDIA RTX 4090)" or "MPS (Apple M2 Pro)"
+  const gpuName = hasGpu
+    ? health.gpu_type!.replace(/^(CUDA|ROCm|MPS|Metal|XPU|DirectML)\s*\((.+)\)$/, '$2') ||
+      health.gpu_type!
+    : null;
+  const gpuBackend = hasGpu ? health.gpu_type!.replace(/\s*\(.+\)$/, '') : null;
+  const isApple = gpuBackend === 'MPS' || gpuBackend === 'Metal';
+  const showBackendVariant = health.backend_variant && health.backend_variant !== 'cpu';
+
+  return (
+    <div className="rounded-lg border border-border/60 p-4">
+      <div className="flex items-center gap-3">
+        {hasGpu ? (
+          isApple ? (
+            <AppleLogo className="h-5 w-5 shrink-0 text-muted-foreground" />
+          ) : (
+            <GpuIcon className="h-5 w-5 shrink-0 text-accent" />
+          )
+        ) : (
+          <Cpu className="h-5 w-5 shrink-0 text-muted-foreground" />
+        )}
+        <div className="flex-1 min-w-0 space-y-0.5">
+          <div className="text-sm font-medium">{hasGpu ? gpuName : 'CPU Only'}</div>
+          <div className="flex flex-wrap items-center gap-x-3 gap-y-1 text-xs text-muted-foreground">
+            {hasGpu ? (
+              <>
+                <span>{gpuBackend}</span>
+                {showBackendVariant && (
+                  <>
+                    <span className="text-border">|</span>
+                    <span className="uppercase">{health.backend_variant}</span>
+                  </>
+                )}
+                {health.vram_used_mb != null && health.vram_used_mb > 0 && (
+                  <>
+                    <span className="text-border">|</span>
+                    <span>{health.vram_used_mb.toFixed(0)} MB VRAM</span>
+                  </>
+                )}
+              </>
+            ) : (
+              <span>No GPU acceleration detected</span>
+            )}
+          </div>
+        </div>
+        {hasGpu && (
+          <div className="flex items-center gap-2 rounded-full border border-accent/30 px-2.5 py-0.5">
+            <span className="relative flex h-1.5 w-1.5">
+              <span className="absolute inline-flex h-full w-full animate-ping rounded-full bg-accent/60" />
+              <span className="relative inline-flex h-1.5 w-1.5 rounded-full bg-accent shadow-[0_0_4px_1px_hsl(var(--accent)/0.4)]" />
+            </span>
+            <span className="text-[10px] font-medium text-muted-foreground">Active</span>
+          </div>
+        )}
+      </div>
+    </div>
+  );
+}
+
+export function GpuPage() {
+  const platform = usePlatform();
+  const queryClient = useQueryClient();
+  const serverUrl = useServerStore((state) => state.serverUrl);
+  const { data: health } = useServerHealth();
+
+  const [restartPhase, setRestartPhase] = useState<RestartPhase>('idle');
+  const [error, setError] = useState<string | null>(null);
+  const [downloadProgress, setDownloadProgress] = useState<CudaDownloadProgress | null>(null);
+  const healthPollRef = useRef<ReturnType<typeof setInterval> | null>(null);
+
+  const {
+    data: cudaStatus,
+    isLoading: _cudaStatusLoading,
+    refetch: refetchCudaStatus,
+  } = useQuery({
+    queryKey: ['cuda-status', serverUrl],
+    queryFn: () => apiClient.getCudaStatus(),
+    refetchInterval: (query) => (query.state.status === 'pending' ? false : 10000),
+    retry: 1,
+    enabled: !!health,
+  });
+
+  const isCurrentlyCuda = health?.backend_variant === 'cuda';
+  const cudaAvailable = cudaStatus?.available ?? false;
+  const cudaDownloading = cudaStatus?.downloading ?? false;
+
+  useEffect(() => {
+    return () => {
+      if (healthPollRef.current) {
+        clearInterval(healthPollRef.current);
+        healthPollRef.current = null;
+      }
+    };
+  }, []);
+
+  useEffect(() => {
+    if (!cudaDownloading || !serverUrl) return;
+
+    const eventSource = new EventSource(`${serverUrl}/backend/cuda-progress`);
+
+    eventSource.onmessage = (event) => {
+      try {
+        const data = JSON.parse(event.data) as CudaDownloadProgress;
+        setDownloadProgress(data);
+
+        if (data.status === 'complete') {
+          eventSource.close();
+          setDownloadProgress(null);
+          refetchCudaStatus();
+        } else if (data.status === 'error') {
+          eventSource.close();
+          setError(data.error || 'Download failed');
+          setDownloadProgress(null);
+          refetchCudaStatus();
+        }
+      } catch (e) {
+        console.error('Error parsing CUDA progress event:', e);
+      }
+    };
+
+    eventSource.onerror = () => {
+      eventSource.close();
+    };
+
+    return () => {
+      eventSource.close();
+    };
+  }, [cudaDownloading, serverUrl, refetchCudaStatus]);
+
+  const clearHealthPolling = useCallback(() => {
+    if (healthPollRef.current) {
+      clearInterval(healthPollRef.current);
+      healthPollRef.current = null;
+    }
+  }, []);
+
+  const startHealthPolling = useCallback(() => {
+    clearHealthPolling();
+
+    healthPollRef.current = setInterval(async () => {
+      try {
+        const result = await apiClient.getHealth();
+        if (result.status === 'healthy') {
+          clearHealthPolling();
+          setRestartPhase('ready');
+          queryClient.invalidateQueries();
+          setTimeout(() => setRestartPhase('idle'), 2000);
+        }
+      } catch {
+        // Server still down, keep polling
+      }
+    }, 1000);
+  }, [queryClient, clearHealthPolling]);
+
+  const restartServerWithPolling = useCallback(
+    async (errorMessage: string) => {
+      setRestartPhase('stopping');
+      try {
+        await platform.lifecycle.restartServer();
+        setRestartPhase('waiting');
+        startHealthPolling();
+      } catch (e: unknown) {
+        clearHealthPolling();
+        setRestartPhase('idle');
+        throw new Error(e instanceof Error ? e.message : errorMessage);
+      }
+    },
+    [platform, startHealthPolling, clearHealthPolling],
+  );
+
+  const handleDownload = async () => {
+    setError(null);
+    try {
+      await apiClient.downloadCudaBackend();
+      refetchCudaStatus();
+    } catch (e: unknown) {
+      const msg = e instanceof Error ? e.message : 'Failed to start download';
+      if (msg.includes('already downloaded')) {
+        refetchCudaStatus();
+      } else {
+        setError(msg);
+      }
+    }
+  };
+
+  const handleRestart = async () => {
+    setError(null);
+    try {
+      await restartServerWithPolling('Restart failed');
+    } catch (e: unknown) {
+      setError(e instanceof Error ? e.message : 'Restart failed');
+    }
+  };
+
+  const handleSwitchToCpu = async () => {
+    setError(null);
+    setRestartPhase('stopping');
+    try {
+      await apiClient.deleteCudaBackend();
+      await restartServerWithPolling('Failed to switch to CPU');
+    } catch (e: unknown) {
+      setError(e instanceof Error ? e.message : 'Failed to switch to CPU');
+      refetchCudaStatus();
+    }
+  };
+
+  const handleDelete = async () => {
+    setError(null);
+    try {
+      await apiClient.deleteCudaBackend();
+      refetchCudaStatus();
+    } catch (e: unknown) {
+      setError(e instanceof Error ? e.message : 'Failed to delete CUDA backend');
+    }
+  };
+
+  const formatBytes = (bytes: number): string => {
+    if (bytes === 0) return '0 B';
+    const k = 1024;
+    const sizes = ['B', 'KB', 'MB', 'GB'];
+    const i = Math.floor(Math.log(bytes) / Math.log(k));
+    return `${(bytes / k ** i).toFixed(1)} ${sizes[i]}`;
+  };
+
+  if (!health) return null;
+
+  const hasNativeGpu =
+    health.gpu_available &&
+    !isCurrentlyCuda &&
+    health.gpu_type &&
+    !health.gpu_type.includes('CUDA');
+
+  return (
+    <div className="space-y-8 max-w-2xl">
+      <GpuInfoCard health={health} />
+
+      {/* CUDA section — only when no native GPU and not already on CUDA */}
+      {!hasNativeGpu && !isCurrentlyCuda && (
+        <SettingSection
+          title="CUDA Backend"
+          description="NVIDIA GPU acceleration via a downloadable CUDA backend."
+        >
+          {/* Download progress */}
+          {cudaDownloading && downloadProgress && (
+            <SettingRow title="Downloading CUDA backend...">
+              <div className="space-y-1.5">
+                <Progress value={downloadProgress.progress} className="h-2" />
+                <div className="flex items-center justify-between text-xs text-muted-foreground">
+                  <span>
+                    {downloadProgress.filename ||
+                      (cudaAvailable ? 'Updating...' : 'Downloading...')}
+                  </span>
+                  <span>
+                    {downloadProgress.total > 0
+                      ? `${formatBytes(downloadProgress.current)} / ${formatBytes(downloadProgress.total)}`
+                      : `${downloadProgress.progress.toFixed(1)}%`}
+                  </span>
+                </div>
+              </div>
+            </SettingRow>
+          )}
+
+          {/* Restart in progress */}
+          {restartPhase !== 'idle' && (
+            <SettingRow
+              title={
+                restartPhase === 'ready'
+                  ? 'Server restarted successfully'
+                  : restartPhase === 'waiting'
+                    ? 'Restarting server...'
+                    : 'Stopping server...'
+              }
+              action={<Loader2 className="h-4 w-4 animate-spin text-muted-foreground" />}
+            />
+          )}
+
+          {/* Error */}
+          {error && (
+            <SettingRow title="Error">
+              <div className="flex items-center gap-2 text-sm text-destructive">
+                <AlertCircle className="h-4 w-4 shrink-0" />
+                <span>{error}</span>
+              </div>
+            </SettingRow>
+          )}
+
+          {/* Actions */}
+          {restartPhase === 'idle' && !cudaDownloading && (
+            <>
+              {!cudaAvailable && !isCurrentlyCuda && (
+                <SettingRow
+                  title="Download CUDA backend"
+                  description="~2.4 GB download. Requires an NVIDIA GPU with CUDA support."
+                  action={
+                    <Button onClick={handleDownload} size="sm">
+                      <Download className="h-3.5 w-3.5 mr-1.5" />
+                      Download
+                    </Button>
+                  }
+                />
+              )}
+
+              {cudaAvailable && !isCurrentlyCuda && platform.metadata.isTauri && (
+                <SettingRow
+                  title="Switch to CUDA backend"
+                  description="CUDA backend is downloaded and ready. Restart to enable."
+                  action={
+                    <Button onClick={handleRestart} size="sm">
+                      <RotateCw className="h-3.5 w-3.5 mr-1.5" />
+                      Restart
+                    </Button>
+                  }
+                />
+              )}
+
+              {isCurrentlyCuda && platform.metadata.isTauri && (
+                <SettingRow
+                  title="Switch to CPU backend"
+                  description="Disable GPU acceleration. You can re-download CUDA later."
+                  action={
+                    <Button onClick={handleSwitchToCpu} variant="outline" size="sm">
+                      <RotateCw className="h-3.5 w-3.5 mr-1.5" />
+                      Switch
+                    </Button>
+                  }
+                />
+              )}
+
+              {cudaAvailable && !isCurrentlyCuda && (
+                <SettingRow
+                  title="Remove CUDA backend"
+                  description="Delete the downloaded CUDA binary to free disk space."
+                  action={
+                    <Button
+                      onClick={handleDelete}
+                      variant="ghost"
+                      size="sm"
+                      className="text-muted-foreground hover:text-destructive"
+                    >
+                      <Trash2 className="h-3.5 w-3.5 mr-1.5" />
+                      Remove
+                    </Button>
+                  }
+                />
+              )}
+            </>
+          )}
+        </SettingSection>
+      )}
+
+      <p className="text-xs text-muted-foreground/60 leading-relaxed">
+        Voicebox automatically detects and uses the best available GPU on your system. On Apple
+        Silicon Macs, the MLX backend runs natively on the Neural Engine and GPU via Metal
+        Performance Shaders (MPS), with no additional setup required. On Windows and Linux with
+        NVIDIA GPUs, you can download an optional CUDA backend for hardware-accelerated inference.
+        AMD ROCm, Intel XPU, and DirectML are also supported where available through PyTorch. When
+        no GPU is detected, Voicebox falls back to CPU — all engines still work, just slower.
+      </p>
+    </div>
+  );
+}

app/src/components/ServerTab/LogsPage.tsx

@@ -0,0 +1,104 @@
+import { useEffect, useRef, useState } from 'react';
+import { Button } from '@/components/ui/button';
+import { cn } from '@/lib/utils/cn';
+import { type LogEntry, useLogStore } from '@/stores/logStore';
+
+function formatTime(timestamp: number): string {
+  const d = new Date(timestamp);
+  return d.toLocaleTimeString(undefined, {
+    hour: '2-digit',
+    minute: '2-digit',
+    second: '2-digit',
+    hour12: false,
+  });
+}
+
+function LogLine({ entry }: { entry: LogEntry }) {
+  return (
+    <div className="flex gap-3 font-mono text-xs leading-5 hover:bg-muted/30">
+      <span className="text-muted-foreground/50 select-none shrink-0">
+        {formatTime(entry.timestamp)}
+      </span>
+      <span
+        className={cn(
+          'whitespace-pre-wrap break-all',
+          entry.stream === 'stderr' ? 'text-orange-400/80' : 'text-muted-foreground',
+        )}
+      >
+        {entry.line}
+      </span>
+    </div>
+  );
+}
+
+export function LogsPage() {
+  const entries = useLogStore((s) => s.entries);
+  const clear = useLogStore((s) => s.clear);
+  const containerRef = useRef<HTMLDivElement>(null);
+  const [autoScroll, setAutoScroll] = useState(true);
+
+  // Auto-scroll to bottom when new entries arrive
+  useEffect(() => {
+    if (autoScroll && containerRef.current) {
+      containerRef.current.scrollTop = containerRef.current.scrollHeight;
+    }
+  }, [entries.length, autoScroll]);
+
+  // Detect manual scroll to disable auto-scroll
+  const handleScroll = () => {
+    const el = containerRef.current;
+    if (!el) return;
+    const atBottom = el.scrollHeight - el.scrollTop - el.clientHeight < 40;
+    setAutoScroll(atBottom);
+  };
+
+  return (
+    <div className="flex flex-col h-full min-h-0">
+      <div className="flex items-center justify-between mb-3">
+        <div>
+          <h3 className="text-sm font-medium">Server Logs</h3>
+          <p className="text-sm text-muted-foreground">
+            {entries.length} {entries.length === 1 ? 'line' : 'lines'}
+          </p>
+        </div>
+        <div className="flex items-center gap-2">
+          {!autoScroll && (
+            <Button
+              variant="outline"
+              size="sm"
+              onClick={() => {
+                setAutoScroll(true);
+                containerRef.current?.scrollTo({ top: containerRef.current.scrollHeight });
+              }}
+            >
+              Scroll to bottom
+            </Button>
+          )}
+          <Button variant="outline" size="sm" onClick={clear}>
+            Clear
+          </Button>
+        </div>
+      </div>
+
+      <div
+        ref={containerRef}
+        onScroll={handleScroll}
+        className="flex-1 min-h-0 overflow-y-auto rounded-md border bg-black/20 p-3"
+      >
+        {entries.length === 0 ? (
+          <div className="text-sm text-muted-foreground/50 font-mono space-y-1">
+            <p>No log output yet.</p>
+            {!import.meta.env?.PROD && (
+              <p>
+                Server logs are only captured when the app manages the server process (production
+                builds).
+              </p>
+            )}
+          </div>
+        ) : (
+          entries.map((entry) => <LogLine key={entry.id} entry={entry} />)
+        )}
+      </div>
+    </div>
+  );
+}

app/src/components/ServerTab/ServerTab.tsx

@@ -1,35 +1,70 @@
-import { ConnectionForm } from '@/components/ServerSettings/ConnectionForm';
+import { Link, Outlet, useMatchRoute } from '@tanstack/react-router';
-import { GenerationSettings } from '@/components/ServerSettings/GenerationSettings';
-import { GpuAcceleration } from '@/components/ServerSettings/GpuAcceleration';
-import { UpdateStatus } from '@/components/ServerSettings/UpdateStatus';
 import { BOTTOM_SAFE_AREA_PADDING } from '@/lib/constants/ui';
 import { cn } from '@/lib/utils/cn';
 import { usePlatform } from '@/platform/PlatformContext';
 import { usePlayerStore } from '@/stores/playerStore';
 
-export function ServerTab() {
+interface SettingsTab {
+  label: string;
+  path:
+    | '/settings'
+    | '/settings/generation'
+    | '/settings/gpu'
+    | '/settings/logs'
+    | '/settings/changelog'
+    | '/settings/about';
+  tauriOnly?: boolean;
+}
+
+const tabs: SettingsTab[] = [
+  { label: 'General', path: '/settings' },
+  { label: 'Generation', path: '/settings/generation' },
+  { label: 'GPU', path: '/settings/gpu', tauriOnly: true },
+  { label: 'Logs', path: '/settings/logs', tauriOnly: true },
+  { label: 'Changelog', path: '/settings/changelog' },
+  { label: 'About', path: '/settings/about' },
+];
+
+export function SettingsLayout() {
   const platform = usePlatform();
   const isPlayerVisible = !!usePlayerStore((state) => state.audioUrl);
+  const matchRoute = useMatchRoute();
+
   return (
-    <div
+    <div className="flex flex-col h-full min-h-0">
-      className={cn('overflow-y-auto flex flex-col', isPlayerVisible && BOTTOM_SAFE_AREA_PADDING)}
+      <nav className="flex gap-1 border-b shrink-0">
-    >
+        {tabs.map((tab) => {
-      <div className="grid gap-4 md:grid-cols-2">
+          if (tab.tauriOnly && !platform.metadata.isTauri) return null;
-        <ConnectionForm />
+
-        <GenerationSettings />
+          const isActive =
-        {platform.metadata.isTauri && <GpuAcceleration />}
+            tab.path === '/settings'
-        {platform.metadata.isTauri && <UpdateStatus />}
+              ? matchRoute({ to: tab.path, fuzzy: false })
-      </div>
+              : matchRoute({ to: tab.path });
-      <div className="py-8 text-center text-sm text-muted-foreground">
+
-        Created by{' '}
+          return (
-        <a
+            <Link
-          href="https://github.com/jamiepine"
+              key={tab.path}
-          target="_blank"
+              to={tab.path}
-          rel="noopener noreferrer"
+              className={cn(
-          className="text-accent hover:underline"
+                'px-4 py-2 text-sm font-medium transition-colors border-b-2 -mb-px',
-        >
+                isActive
-          Jamie Pine
+                  ? 'border-accent text-foreground'
-        </a>
+                  : 'border-transparent text-muted-foreground hover:text-foreground hover:border-muted-foreground/30',
+              )}
+            >
+              {tab.label}
+            </Link>
+          );
+        })}
+      </nav>
+
+      <div
+        className={cn(
+          'flex-1 overflow-y-auto pt-6 pb-6 px-2 -mx-2',
+          isPlayerVisible && BOTTOM_SAFE_AREA_PADDING,
+        )}
+      >
+        <Outlet />
       </div>
     </div>
   );

app/src/components/ServerTab/SettingRow.tsx

@@ -0,0 +1,62 @@
+import type { ReactNode } from 'react';
+
+/**
+ * A section header with title and optional description, separated by a border.
+ */
+export function SettingSection({
+  title,
+  description,
+  children,
+}: {
+  title?: string;
+  description?: string;
+  children: ReactNode;
+}) {
+  return (
+    <div className="space-y-1">
+      {title && <h3 className="text-sm font-medium">{title}</h3>}
+      {description && <p className="text-sm text-muted-foreground">{description}</p>}
+      <div className={`${title || description ? 'pt-3' : ''} space-y-0 divide-y divide-border/60`}>
+        {children}
+      </div>
+    </div>
+  );
+}
+
+/**
+ * A single settings row: label+description on the left, action on the right.
+ * Use for toggles, inputs, buttons, badges — any control type.
+ */
+export function SettingRow({
+  title,
+  description,
+  htmlFor,
+  action,
+  children,
+}: {
+  title: string;
+  description?: string;
+  htmlFor?: string;
+  /** Right-aligned control (checkbox, button, badge, etc.) */
+  action?: ReactNode;
+  /** Full-width content rendered below the label row (for sliders, inputs, etc.) */
+  children?: ReactNode;
+}) {
+  return (
+    <div className="py-3">
+      <div className="flex items-center justify-between gap-8">
+        <div className="min-w-0">
+          <label
+            htmlFor={htmlFor}
+            className={`text-sm font-medium leading-none select-none ${htmlFor ? 'cursor-pointer' : ''}`}
+          >
+            {title}
+          </label>
+          {description && <p className="text-sm text-muted-foreground mt-0.5">{description}</p>}
+        </div>
+        {action && <div className="shrink-0">{action}</div>}
+      </div>
+      {children && <div className="mt-3">{children}</div>}
+    </div>
+  );
+}

app/src/components/Sidebar.tsx

@@ -1,5 +1,5 @@
 import { Link, useMatchRoute } from '@tanstack/react-router';
-import { AudioLines, Box, Mic, Server, Speaker, Volume2, Wand2 } from 'lucide-react';
+import { AudioLines, Box, Mic, Settings, Speaker, Volume2, Wand2 } from 'lucide-react';
 import { useEffect, useState } from 'react';
 import voiceboxLogo from '@/assets/voicebox-logo.png';
 import { cn } from '@/lib/utils/cn';
@@ -19,7 +19,7 @@ const tabs = [
   { id: 'effects', path: '/effects', icon: Wand2, label: 'Effects' },
   { id: 'audio', path: '/audio', icon: Speaker, label: 'Audio' },
   { id: 'models', path: '/models', icon: Box, label: 'Models' },
-  { id: 'server', path: '/server', icon: Server, label: 'Server' },
+  { id: 'settings', path: '/settings', icon: Settings, label: 'Settings' },
 ];
 
 export function Sidebar({ isMacOS }: SidebarProps) {
@@ -54,9 +54,10 @@ export function Sidebar({ isMacOS }: SidebarProps) {
       <div className="flex flex-col gap-3">
         {tabs.map((tab, index) => {
           const Icon = tab.icon;
-          // For index route, use exact match; for others, use default matching
           const isActive =
-            tab.path === '/' ? matchRoute({ to: '/', exact: true }) : matchRoute({ to: tab.path });
+            tab.path === '/'
+              ? matchRoute({ to: '/', fuzzy: false })
+              : matchRoute({ to: tab.path, fuzzy: true });
 
           // Accent fades as buttons get further from the logo
           const accentOpacity = Math.max(0.08, 0.5 - index * 0.07);
@@ -98,7 +99,7 @@ export function Sidebar({ isMacOS }: SidebarProps) {
         <span className="text-[10px] text-muted-foreground/50">v{version}</span>
         {updateStatus.available && (
           <Link
-            to="/server"
+            to="/settings"
             className="text-[9px] font-semibold tracking-wide uppercase px-2 py-0.5 rounded-full bg-accent/15 text-accent hover:bg-accent/25 transition-colors"
           >
             Update

app/src/components/ui/slider.tsx

@@ -1,5 +1,5 @@
-import * as React from 'react';
 import * as SliderPrimitive from '@radix-ui/react-slider';
+import * as React from 'react';
 import { cn } from '@/lib/utils/cn';
 
 const Slider = React.forwardRef<
@@ -14,7 +14,7 @@ const Slider = React.forwardRef<
     <SliderPrimitive.Track className="relative h-2 w-full grow overflow-hidden rounded-full bg-secondary">
       <SliderPrimitive.Range className="absolute h-full bg-primary" />
     </SliderPrimitive.Track>
-    <SliderPrimitive.Thumb className="block h-5 w-5 rounded-full border-2 border-primary bg-background ring-offset-background transition-colors focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring focus-visible:ring-offset-2 disabled:pointer-events-none disabled:opacity-50 translate-x-0.5" />
+    <SliderPrimitive.Thumb className="block h-0 w-0 outline-none disabled:pointer-events-none disabled:opacity-50 after:block after:h-5 after:w-5 after:rounded-full after:border-2 after:border-primary after:bg-background after:ring-offset-background after:transition-colors after:absolute after:top-1/2 after:left-1/2 after:-translate-x-1/2 after:-translate-y-1/2 focus-visible:after:ring-2 focus-visible:after:ring-ring focus-visible:after:ring-offset-2" />
   </SliderPrimitive.Root>
 ));
 Slider.displayName = SliderPrimitive.Root.displayName;

app/src/components/ui/toaster.tsx

@@ -1,3 +1,4 @@
+import { usePlayerStore } from '@/stores/playerStore';
 import {
   Toast,
   ToastClose,
@@ -10,6 +11,7 @@ import { useToast } from './use-toast';
 
 export function Toaster() {
   const { toasts } = useToast();
+  const isPlayerOpen = !!usePlayerStore((s) => s.audioUrl);
 
   return (
     <ToastProvider>
@@ -23,7 +25,7 @@ export function Toaster() {
           <ToastClose />
         </Toast>
       ))}
-      <ToastViewport />
+      <ToastViewport className={isPlayerOpen ? 'sm:bottom-44' : ''} />
     </ToastProvider>
   );
 }

app/src/components/ui/toggle.tsx

@@ -0,0 +1,48 @@
+import * as React from 'react';
+import { cn } from '@/lib/utils/cn';
+
+export interface ToggleProps {
+  checked?: boolean;
+  onCheckedChange?: (checked: boolean) => void;
+  disabled?: boolean;
+  className?: string;
+  id?: string;
+}
+
+const Toggle = React.forwardRef<HTMLButtonElement, ToggleProps>(
+  ({ checked = false, onCheckedChange, disabled = false, className, id, ...props }, ref) => {
+    return (
+      <button
+        type="button"
+        ref={ref}
+        id={id}
+        role="switch"
+        aria-checked={checked}
+        disabled={disabled}
+        onClick={() => {
+          if (!disabled && onCheckedChange) {
+            onCheckedChange(!checked);
+          }
+        }}
+        className={cn(
+          'relative inline-flex h-5 w-9 shrink-0 items-center rounded-full transition-colors',
+          'focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring focus-visible:ring-offset-2',
+          checked ? 'bg-accent' : 'bg-muted-foreground/25',
+          disabled ? 'opacity-50 cursor-not-allowed' : 'cursor-pointer',
+          className,
+        )}
+        {...props}
+      >
+        <span
+          className={cn(
+            'pointer-events-none block h-4 w-4 rounded-full bg-white shadow-sm transition-transform',
+            checked ? 'translate-x-[18px]' : 'translate-x-[2px]',
+          )}
+        />
+      </button>
+    );
+  },
+);
+Toggle.displayName = 'Toggle';
+
+export { Toggle };

app/src/global.d.ts

@@ -1,3 +1,8 @@
 interface Window {
   __voiceboxServerStartedByApp?: boolean;
 }
+
+declare module 'virtual:changelog' {
+  const raw: string;
+  export default raw;
+}

app/src/lib/api/client.ts

@@ -32,8 +32,24 @@ import type {
   TranscriptionResponse,
   VoiceProfileCreate,
   VoiceProfileResponse,
+  WhisperModelSize,
 } from './types';
 
+function formatErrorDetail(detail: unknown, fallback: string): string {
+  if (typeof detail === 'string') return detail;
+  if (Array.isArray(detail)) {
+    return detail
+      .map((e: Record<string, unknown>) => e.msg || e.message || JSON.stringify(e))
+      .join('; ');
+  }
+  if (detail && typeof detail === 'object') {
+    const obj = detail as Record<string, unknown>;
+    if (typeof obj.message === 'string') return obj.message;
+    return JSON.stringify(detail);
+  }
+  return fallback;
+}
+
 class ApiClient {
   private getBaseUrl(): string {
     const serverUrl = useServerStore.getState().serverUrl;
@@ -54,7 +70,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.json();
@@ -113,7 +129,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.json();
@@ -147,7 +163,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.blob();
@@ -167,7 +183,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.json();
@@ -187,7 +203,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.json();
@@ -257,7 +273,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.blob();
@@ -271,7 +287,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.blob();
@@ -297,7 +313,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.json();
@@ -318,12 +334,19 @@ class ApiClient {
   }
 
   // Transcription
-  async transcribeAudio(file: File, language?: LanguageCode): Promise<TranscriptionResponse> {
+  async transcribeAudio(
+    file: File,
+    language?: LanguageCode,
+    model?: WhisperModelSize,
+  ): Promise<TranscriptionResponse> {
     const formData = new FormData();
     formData.append('file', file);
     if (language) {
       formData.append('language', language);
     }
+    if (model) {
+      formData.append('model', model);
+    }
 
     const url = `${this.getBaseUrl()}/transcribe`;
     const response = await fetch(url, {
@@ -335,7 +358,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.json();
@@ -608,7 +631,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.blob();
@@ -705,7 +728,7 @@ class ApiClient {
       const error = await response.json().catch(() => ({
         detail: response.statusText,
       }));
-      throw new Error(error.detail || `HTTP error! status: ${response.status}`);
+      throw new Error(formatErrorDetail(error.detail, `HTTP error! status: ${response.status}`));
     }
 
     return response.blob();

app/src/lib/api/types.ts

@@ -42,8 +42,8 @@ export interface GenerationRequest {
   text: string;
   language: LanguageCode;
   seed?: number;
-  model_size?: '1.7B' | '0.6B';
+  model_size?: '1.7B' | '0.6B' | '1B' | '3B';
-  engine?: 'qwen' | 'luxtts' | 'chatterbox' | 'chatterbox_turbo';
+  engine?: 'qwen' | 'luxtts' | 'chatterbox' | 'chatterbox_turbo' | 'tada';
   instruct?: string;
   max_chunk_chars?: number;
   crossfade_ms?: number;
@@ -99,8 +99,11 @@ export interface HistoryListResponse {
   total: number;
 }
 
+export type WhisperModelSize = 'base' | 'small' | 'medium' | 'large' | 'turbo';
+
 export interface TranscriptionRequest {
   language?: LanguageCode;
+  model?: WhisperModelSize;
 }
 
 export interface TranscriptionResponse {

app/src/lib/constants/languages.ts

@@ -66,6 +66,7 @@ export const ENGINE_LANGUAGES: Record<string, readonly LanguageCode[]> = {
     'zh',
   ],
   chatterbox_turbo: ['en'],
+  tada: ['en', 'ar', 'zh', 'de', 'es', 'fr', 'it', 'ja', 'pl', 'pt'],
 } as const;
 
 /** Helper: get language options for a given engine. */

app/src/lib/hooks/useGenerationForm.ts

@@ -15,9 +15,9 @@ const generationSchema = z.object({
   text: z.string().min(1, '').max(50000),
   language: z.enum(LANGUAGE_CODES as [LanguageCode, ...LanguageCode[]]),
   seed: z.number().int().optional(),
-  modelSize: z.enum(['1.7B', '0.6B']).optional(),
+  modelSize: z.enum(['1.7B', '0.6B', '1B', '3B']).optional(),
   instruct: z.string().max(500).optional(),
-  engine: z.enum(['qwen', 'luxtts', 'chatterbox', 'chatterbox_turbo']).optional(),
+  engine: z.enum(['qwen', 'luxtts', 'chatterbox', 'chatterbox_turbo', 'tada']).optional(),
 });
 
 export type GenerationFormValues = z.infer<typeof generationSchema>;
@@ -79,7 +79,11 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
             ? 'chatterbox-tts'
             : engine === 'chatterbox_turbo'
               ? 'chatterbox-turbo'
-              : `qwen-tts-${data.modelSize}`;
+              : engine === 'tada'
+                ? data.modelSize === '3B'
+                  ? 'tada-3b-ml'
+                  : 'tada-1b'
+                : `qwen-tts-${data.modelSize}`;
       const displayName =
         engine === 'luxtts'
           ? 'LuxTTS'
@@ -87,9 +91,13 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
             ? 'Chatterbox TTS'
             : engine === 'chatterbox_turbo'
               ? 'Chatterbox Turbo'
-              : data.modelSize === '1.7B'
+              : engine === 'tada'
-                ? 'Qwen TTS 1.7B'
+                ? data.modelSize === '3B'
-                : 'Qwen TTS 0.6B';
+                  ? 'TADA 3B Multilingual'
+                  : 'TADA 1B'
+                : data.modelSize === '1.7B'
+                  ? 'Qwen TTS 1.7B'
+                  : 'Qwen TTS 0.6B';
 
       // Check if model needs downloading
       try {
@@ -104,7 +112,7 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
         console.error('Failed to check model status:', error);
       }
 
-      const isQwen = engine === 'qwen';
+      const hasModelSizes = engine === 'qwen' || engine === 'tada';
       const effectsChain = options.getEffectsChain?.();
       // This now returns immediately with status="generating"
       const result = await generation.mutateAsync({
@@ -112,9 +120,9 @@ export function useGenerationForm(options: UseGenerationFormOptions = {}) {
         text: data.text,
         language: data.language,
         seed: data.seed,
-        model_size: isQwen ? data.modelSize : undefined,
+        model_size: hasModelSizes ? data.modelSize : undefined,
         engine,
-        instruct: isQwen ? data.instruct || undefined : undefined,
+        instruct: engine === 'qwen' ? data.instruct || undefined : undefined,
         max_chunk_chars: maxChunkChars,
         crossfade_ms: crossfadeMs,
         normalize: normalizeAudio,

app/src/lib/hooks/useGenerationProgress.ts

@@ -75,8 +75,8 @@ export function useGenerationProgress() {
             currentSources.delete(id);
             removePendingGeneration(id);
 
-            // Refresh history to pick up the completed generation
+            // Refetch history to pick up the completed generation
-            queryClient.invalidateQueries({ queryKey: ['history'] });
+            queryClient.refetchQueries({ queryKey: ['history'] });
 
             // If this generation was queued for a story, add it now
             const storyId = removePendingStoryAdd(id);
@@ -120,7 +120,7 @@ export function useGenerationProgress() {
             removePendingGeneration(id);
             removePendingStoryAdd(id);
 
-            queryClient.invalidateQueries({ queryKey: ['history'] });
+            queryClient.refetchQueries({ queryKey: ['history'] });
 
             toast({
               title: data.status === 'not_found' ? 'Generation not found' : 'Generation failed',
@@ -134,11 +134,12 @@ export function useGenerationProgress() {
       };
 
       source.onerror = () => {
-        // EventSource auto-reconnects, but if we get repeated errors
+        // SSE connection dropped — clean up and refresh history so any
-        // just clean up
+        // completed/failed generation still appears in the list
         source.close();
         currentSources.delete(id);
         removePendingGeneration(id);
+        queryClient.refetchQueries({ queryKey: ['history'] });
       };
 
       currentSources.set(id, source);

app/src/lib/hooks/useSystemAudioCapture.ts

@@ -1,4 +1,4 @@
-import { useState, useRef, useCallback, useEffect } from 'react';
+import { useCallback, useEffect, useRef, useState } from 'react';
 import { usePlatform } from '@/platform/PlatformContext';
 
 interface UseSystemAudioCaptureOptions {
@@ -100,9 +100,7 @@ export function useSystemAudioCapture({
       onRecordingComplete?.(blob, recordedDuration);
     } catch (err) {
       const errorMessage =
-        err instanceof Error
+        err instanceof Error ? err.message : 'Failed to stop system audio capture.';
-          ? err.message
-          : 'Failed to stop system audio capture.';
       setError(errorMessage);
     }
   }, [isRecording, onRecordingComplete, platform]);

app/src/lib/hooks/useTranscription.ts

@@ -1,10 +1,18 @@
 import { useMutation } from '@tanstack/react-query';
 import { apiClient } from '@/lib/api/client';
+import type { WhisperModelSize } from '@/lib/api/types';
 import type { LanguageCode } from '@/lib/constants/languages';
 
 export function useTranscription() {
   return useMutation({
-    mutationFn: ({ file, language }: { file: File; language?: LanguageCode }) =>
+    mutationFn: ({
-      apiClient.transcribeAudio(file, language),
+      file,
+      language,
+      model,
+    }: {
+      file: File;
+      language?: LanguageCode;
+      model?: WhisperModelSize;
+    }) => apiClient.transcribeAudio(file, language, model),
   });
 }

app/src/lib/utils/parseChangelog.ts

@@ -0,0 +1,37 @@
+export interface ChangelogEntry {
+  version: string;
+  date: string | null;
+  body: string;
+}
+
+/**
+ * Parses a Keep-a-Changelog style markdown string into structured entries.
+ *
+ * Splits on `## [version]` headings and extracts the version + date from each.
+ * The body is the raw markdown between headings (trimmed), with the leading
+ * `# Changelog` title and trailing link references stripped.
+ */
+export function parseChangelog(raw: string): ChangelogEntry[] {
+  const entries: ChangelogEntry[] = [];
+
+  // Strip trailing link reference definitions (e.g. [0.1.0]: https://...)
+  const cleaned = raw.replace(/^\[[\w.]+\]:.*$/gm, '').trimEnd();
+
+  // Match `## [version]` or `## [version] - date`
+  const headingRe = /^## \[(.+?)\](?:\s*-\s*(.+))?$/gm;
+  const matches = [...cleaned.matchAll(headingRe)];
+
+  for (let i = 0; i < matches.length; i++) {
+    const match = matches[i];
+    const version = match[1];
+    const date = match[2]?.trim() || null;
+
+    const start = match.index! + match[0].length;
+    const end = i + 1 < matches.length ? matches[i + 1].index! : cleaned.length;
+    const body = cleaned.slice(start, end).trim();
+
+    entries.push({ version, date, body });
+  }
+
+  return entries;
+}

app/src/platform/types.ts

@@ -50,12 +50,18 @@ export interface PlatformAudio {
   stopPlayback(): void;
 }
 
+export interface ServerLogEntry {
+  stream: 'stdout' | 'stderr';
+  line: string;
+}
+
 export interface PlatformLifecycle {
   startServer(remote?: boolean, modelsDir?: string | null): Promise<string>;
   stopServer(): Promise<void>;
   restartServer(modelsDir?: string | null): Promise<string>;
   setKeepServerRunning(keep: boolean): Promise<void>;
   setupWindowCloseHandler(): Promise<void>;
+  subscribeToServerLogs(callback: (entry: ServerLogEntry) => void): () => void;
   onServerReady?: () => void;
 }
 

app/src/router.tsx

@@ -1,10 +1,22 @@
-import { createRootRoute, createRoute, createRouter, Outlet } from '@tanstack/react-router';
+import {
+  createRootRoute,
+  createRoute,
+  createRouter,
+  Outlet,
+  redirect,
+} from '@tanstack/react-router';
 import { AppFrame } from '@/components/AppFrame/AppFrame';
 import { AudioTab } from '@/components/AudioTab/AudioTab';
 import { EffectsTab } from '@/components/EffectsTab/EffectsTab';
 import { MainEditor } from '@/components/MainEditor/MainEditor';
 import { ModelsTab } from '@/components/ModelsTab/ModelsTab';
-import { ServerTab } from '@/components/ServerTab/ServerTab';
+import { AboutPage } from '@/components/ServerTab/AboutPage';
+import { ChangelogPage } from '@/components/ServerTab/ChangelogPage';
+import { GeneralPage } from '@/components/ServerTab/GeneralPage';
+import { GenerationPage } from '@/components/ServerTab/GenerationPage';
+import { GpuPage } from '@/components/ServerTab/GpuPage';
+import { LogsPage } from '@/components/ServerTab/LogsPage';
+import { SettingsLayout } from '@/components/ServerTab/ServerTab';
 import { Sidebar } from '@/components/Sidebar';
 import { StoriesTab } from '@/components/StoriesTab/StoriesTab';
 import { Toaster } from '@/components/ui/toaster';
@@ -120,11 +132,57 @@ const modelsRoute = createRoute({
   component: ModelsTab,
 });
 
-// Server route
+// Settings layout route (parent for sub-tabs)
-const serverRoute = createRoute({
+const settingsRoute = createRoute({
+  getParentRoute: () => rootRoute,
+  path: '/settings',
+  component: SettingsLayout,
+});
+
+// Settings sub-routes
+const settingsGeneralRoute = createRoute({
+  getParentRoute: () => settingsRoute,
+  path: '/',
+  component: GeneralPage,
+});
+
+const settingsGenerationRoute = createRoute({
+  getParentRoute: () => settingsRoute,
+  path: '/generation',
+  component: GenerationPage,
+});
+
+const settingsGpuRoute = createRoute({
+  getParentRoute: () => settingsRoute,
+  path: '/gpu',
+  component: GpuPage,
+});
+
+const settingsChangelogRoute = createRoute({
+  getParentRoute: () => settingsRoute,
+  path: '/changelog',
+  component: ChangelogPage,
+});
+
+const settingsLogsRoute = createRoute({
+  getParentRoute: () => settingsRoute,
+  path: '/logs',
+  component: LogsPage,
+});
+
+const settingsAboutRoute = createRoute({
+  getParentRoute: () => settingsRoute,
+  path: '/about',
+  component: AboutPage,
+});
+
+// Redirect old /server path to /settings
+const serverRedirectRoute = createRoute({
   getParentRoute: () => rootRoute,
   path: '/server',
-  component: ServerTab,
+  beforeLoad: () => {
+    throw redirect({ to: '/settings' });
+  },
 });
 
 // Route tree
@@ -135,7 +193,15 @@ const routeTree = rootRoute.addChildren([
   audioRoute,
   effectsRoute,
   modelsRoute,
-  serverRoute,
+  settingsRoute.addChildren([
+    settingsGeneralRoute,
+    settingsGenerationRoute,
+    settingsGpuRoute,
+    settingsLogsRoute,
+    settingsChangelogRoute,
+    settingsAboutRoute,
+  ]),
+  serverRedirectRoute,
 ]);
 
 // Create router

app/src/stores/logStore.ts

@@ -0,0 +1,31 @@
+import { create } from 'zustand';
+import type { ServerLogEntry } from '@/platform/types';
+
+const MAX_LOG_ENTRIES = 2000;
+
+let nextLogEntryId = 0;
+
+export interface LogEntry extends ServerLogEntry {
+  id: number;
+  timestamp: number;
+}
+
+interface LogStore {
+  entries: LogEntry[];
+  addEntry: (entry: ServerLogEntry) => void;
+  clear: () => void;
+}
+
+export const useLogStore = create<LogStore>((set) => ({
+  entries: [],
+  addEntry: (entry) =>
+    set((state) => {
+      const newEntry: LogEntry = { ...entry, id: nextLogEntryId++, timestamp: Date.now() };
+      const entries = [...state.entries, newEntry];
+      if (entries.length > MAX_LOG_ENTRIES) {
+        return { entries: entries.slice(entries.length - MAX_LOG_ENTRIES) };
+      }
+      return { entries };
+    }),
+  clear: () => set({ entries: [] }),
+}));

app/tsconfig.node.json

@@ -6,5 +6,5 @@
     "moduleResolution": "bundler",
     "allowSyntheticDefaultImports": true
   },
-  "include": ["vite.config.ts"]
+  "include": ["vite.config.ts", "plugins/**/*.ts"]
 }

app/vite.config.ts

@@ -2,9 +2,10 @@ import path from 'node:path';
 import tailwindcss from '@tailwindcss/vite';
 import react from '@vitejs/plugin-react';
 import { defineConfig } from 'vite';
+import { changelogPlugin } from './plugins/changelog';
 
 export default defineConfig({
-  plugins: [tailwindcss(), react()],
+  plugins: [tailwindcss(), react(), changelogPlugin(path.resolve(__dirname, '..'))],
   resolve: {
     alias: {
       '@': path.resolve(__dirname, './src'),

backend/__init__.py

@@ -1,3 +1,3 @@
 # Backend package
 
-__version__ = "0.2.3"
+__version__ = "0.3.1"

backend/app.py

@@ -77,6 +77,7 @@ def create_app() -> FastAPI:
     _configure_cors(application)
     register_routers(application)
     _register_lifecycle(application)
+    _mount_frontend(application)
 
     return application
 
@@ -104,6 +105,43 @@ def _configure_cors(application: FastAPI) -> None:
     )
 
 
+def _mount_frontend(application: FastAPI) -> None:
+    """Serve the built web frontend when present (Docker / web deployment).
+
+    The Dockerfile copies the Vite build output to ``/app/frontend/``.  When
+    that directory exists we mount static assets and add a catch-all route so
+    the React SPA handles client-side routing.  In dev or API-only mode the
+    directory is absent and this function is a no-op.
+    """
+    frontend_dir = Path(__file__).resolve().parent.parent / "frontend"
+    if not frontend_dir.is_dir():
+        return
+
+    from fastapi.staticfiles import StaticFiles
+    from fastapi.responses import FileResponse
+
+    # Mount hashed assets (JS, CSS, images) that Vite places under /assets
+    assets_dir = frontend_dir / "assets"
+    if assets_dir.is_dir():
+        application.mount(
+            "/assets",
+            StaticFiles(directory=str(assets_dir)),
+            name="frontend-assets",
+        )
+
+    # SPA catch-all: serve files if they exist, otherwise index.html for
+    # client-side routes like /voices, /stories, /models, etc.
+    @application.get("/{full_path:path}")
+    async def serve_spa(full_path: str):
+        file_path = (frontend_dir / full_path).resolve()
+        # Guard against path traversal — only serve files inside frontend_dir
+        if full_path and file_path.is_file() and str(file_path).startswith(str(frontend_dir)):
+            return FileResponse(file_path)
+        return FileResponse(frontend_dir / "index.html", media_type="text/html")
+
+    logger.info("Frontend: serving SPA from %s", frontend_dir)
+
+
 def _get_gpu_status() -> str:
     """Return a human-readable string describing GPU availability."""
     backend_type = get_backend_type()

backend/backends/__init__.py

@@ -134,6 +134,7 @@ class STTBackend(Protocol):
         self,
         audio_path: str,
         language: Optional[str] = None,
+        model_size: Optional[str] = None,
     ) -> str:
         """
         Transcribe audio to text.
@@ -165,6 +166,7 @@ TTS_ENGINES = {
     "luxtts": "LuxTTS",
     "chatterbox": "Chatterbox TTS",
     "chatterbox_turbo": "Chatterbox Turbo",
+    "tada": "TADA",
 }
 
 
@@ -258,6 +260,24 @@ def _get_non_qwen_tts_configs() -> list[ModelConfig]:
             needs_trim=True,
             languages=["en"],
         ),
+        ModelConfig(
+            model_name="tada-1b",
+            display_name="TADA 1B (English)",
+            engine="tada",
+            hf_repo_id="HumeAI/tada-1b",
+            model_size="1B",
+            size_mb=4000,
+            languages=["en"],
+        ),
+        ModelConfig(
+            model_name="tada-3b-ml",
+            display_name="TADA 3B Multilingual",
+            engine="tada",
+            hf_repo_id="HumeAI/tada-3b-ml",
+            model_size="3B",
+            size_mb=8000,
+            languages=["en", "ar", "zh", "de", "es", "fr", "it", "ja", "pl", "pt"],
+        ),
     ]
 
 
@@ -338,10 +358,12 @@ def engine_has_model_sizes(engine: str) -> bool:
 
 
 async def load_engine_model(engine: str, model_size: str = "default") -> None:
-    """Load a model for the given engine, handling the Qwen model_size special case."""
+    """Load a model for the given engine, handling engines with multiple model sizes."""
     backend = get_tts_backend_for_engine(engine)
     if engine == "qwen":
         await backend.load_model_async(model_size)
+    elif engine == "tada":
+        await backend.load_model(model_size)
     else:
         await backend.load_model()
 
@@ -357,7 +379,7 @@ async def ensure_model_cached_or_raise(engine: str, model_size: str = "default")
             cfg = c
             break
 
-    if engine == "qwen":
+    if engine in ("qwen", "tada"):
         if not backend._is_model_cached(model_size):
             raise HTTPException(
                 status_code=400,
@@ -489,6 +511,10 @@ def get_tts_backend_for_engine(engine: str) -> TTSBackend:
             from .chatterbox_turbo_backend import ChatterboxTurboTTSBackend
 
             backend = ChatterboxTurboTTSBackend()
+        elif engine == "tada":
+            from .hume_backend import HumeTadaBackend
+
+            backend = HumeTadaBackend()
         else:
             raise ValueError(f"Unknown TTS engine: {engine}. Supported: {list(TTS_ENGINES.keys())}")
 

backend/backends/hume_backend.py

@@ -0,0 +1,347 @@
+"""
+HumeAI TADA TTS backend implementation.
+
+Wraps HumeAI's TADA (Text-Acoustic Dual Alignment) model for
+high-quality voice cloning. Two model variants:
+  - tada-1b: English-only, ~2B params (Llama 3.2 1B base)
+  - tada-3b-ml: Multilingual, ~4B params (Llama 3.2 3B base)
+
+Both use a shared encoder/codec (HumeAI/tada-codec). The encoder
+produces 1:1 aligned token embeddings from reference audio, and the
+causal LM generates speech via flow-matching diffusion.
+
+24kHz output, bf16 inference on CUDA, fp32 on CPU.
+"""
+
+import asyncio
+import logging
+import threading
+from typing import ClassVar, List, Optional, Tuple
+
+import numpy as np
+
+from . import TTSBackend
+from .base import (
+    is_model_cached,
+    get_torch_device,
+    combine_voice_prompts as _combine_voice_prompts,
+    model_load_progress,
+)
+from ..utils.cache import get_cache_key, get_cached_voice_prompt, cache_voice_prompt
+
+logger = logging.getLogger(__name__)
+
+# HuggingFace repos
+TADA_CODEC_REPO = "HumeAI/tada-codec"
+TADA_1B_REPO = "HumeAI/tada-1b"
+TADA_3B_ML_REPO = "HumeAI/tada-3b-ml"
+
+TADA_MODEL_REPOS = {
+    "1B": TADA_1B_REPO,
+    "3B": TADA_3B_ML_REPO,
+}
+
+# Key weight files for cache detection
+_TADA_MODEL_WEIGHT_FILES = [
+    "model.safetensors",
+]
+
+_TADA_CODEC_WEIGHT_FILES = [
+    "encoder/model.safetensors",
+]
+
+
+class HumeTadaBackend:
+    """HumeAI TADA TTS backend for high-quality voice cloning."""
+
+    _load_lock: ClassVar[threading.Lock] = threading.Lock()
+
+    def __init__(self):
+        self.model = None
+        self.encoder = None
+        self.model_size = "1B"  # default to 1B
+        self._device = None
+        self._model_load_lock = asyncio.Lock()
+
+    def _get_device(self) -> str:
+        # Force CPU on macOS — MPS has issues with flow matching
+        # and large vocab lm_head (>65536 output channels)
+        return get_torch_device(force_cpu_on_mac=True)
+
+    def is_loaded(self) -> bool:
+        return self.model is not None
+
+    def _get_model_path(self, model_size: str = "1B") -> str:
+        return TADA_MODEL_REPOS.get(model_size, TADA_1B_REPO)
+
+    def _is_model_cached(self, model_size: str = "1B") -> bool:
+        repo = TADA_MODEL_REPOS.get(model_size, TADA_1B_REPO)
+        model_cached = is_model_cached(repo, required_files=_TADA_MODEL_WEIGHT_FILES)
+        codec_cached = is_model_cached(TADA_CODEC_REPO, required_files=_TADA_CODEC_WEIGHT_FILES)
+        return model_cached and codec_cached
+
+    async def load_model(self, model_size: str = "1B") -> None:
+        """Load the TADA model and encoder."""
+        if self.model is not None and self.model_size == model_size:
+            return
+        async with self._model_load_lock:
+            if self.model is not None and self.model_size == model_size:
+                return
+            # Unload existing model if switching sizes
+            if self.model is not None:
+                self.unload_model()
+            self.model_size = model_size
+            await asyncio.to_thread(self._load_model_sync, model_size)
+
+    def _load_model_sync(self, model_size: str = "1B"):
+        """Synchronous model loading with progress tracking."""
+        model_name = f"tada-{model_size.lower()}"
+        is_cached = self._is_model_cached(model_size)
+        repo = TADA_MODEL_REPOS.get(model_size, TADA_1B_REPO)
+
+        with model_load_progress(model_name, is_cached):
+            # Install DAC shim before importing tada — tada's encoder/decoder
+            # import dac.nn.layers.Snake1d which requires the descript-audio-codec
+            # package.  The real package pulls in onnx/tensorboard/matplotlib via
+            # descript-audiotools, so we use a lightweight shim instead.
+            from ..utils.dac_shim import install_dac_shim
+            install_dac_shim()
+
+            import torch
+            from huggingface_hub import snapshot_download
+
+            device = self._get_device()
+            self._device = device
+            logger.info(f"Loading HumeAI TADA {model_size} on {device}...")
+
+            # Download codec (encoder + decoder) if not cached
+            logger.info("Downloading TADA codec...")
+            snapshot_download(
+                repo_id=TADA_CODEC_REPO,
+                token=None,
+                allow_patterns=["*.safetensors", "*.json", "*.txt", "*.bin"],
+            )
+
+            # Download model weights if not cached
+            logger.info(f"Downloading TADA {model_size} model...")
+            snapshot_download(
+                repo_id=repo,
+                token=None,
+                allow_patterns=["*.safetensors", "*.json", "*.txt", "*.bin", "*.model"],
+            )
+
+            # TADA hardcodes "meta-llama/Llama-3.2-1B" as the tokenizer
+            # source in its Aligner and TadaForCausalLM.from_pretrained().
+            # That repo is gated (requires Meta license acceptance).
+            # Download the tokenizer from an ungated mirror and get its
+            # local cache path so we can point TADA at it directly.
+            logger.info("Downloading Llama tokenizer (ungated mirror)...")
+            tokenizer_path = snapshot_download(
+                repo_id="unsloth/Llama-3.2-1B",
+                token=None,
+                allow_patterns=["tokenizer*", "special_tokens*"],
+            )
+
+            # Determine dtype — use bf16 on CUDA for ~50% memory savings
+            if device == "cuda" and torch.cuda.is_bf16_supported():
+                model_dtype = torch.bfloat16
+            else:
+                model_dtype = torch.float32
+
+            # Patch the Aligner config class to use the local tokenizer
+            # path instead of the gated "meta-llama/Llama-3.2-1B" default.
+            # This avoids monkey-patching AutoTokenizer.from_pretrained
+            # which corrupts the classmethod descriptor for other engines.
+            from tada.modules.aligner import AlignerConfig
+            AlignerConfig.tokenizer_name = tokenizer_path
+
+            # Load encoder (only needed for voice prompt encoding)
+            from tada.modules.encoder import Encoder
+            logger.info("Loading TADA encoder...")
+            self.encoder = Encoder.from_pretrained(
+                TADA_CODEC_REPO, subfolder="encoder"
+            ).to(device)
+            self.encoder.eval()
+
+            # Load the causal LM (includes decoder for wav generation).
+            # TadaForCausalLM.from_pretrained() calls
+            #   getattr(config, "tokenizer_name", "meta-llama/Llama-3.2-1B")
+            # which hits the gated repo. Pre-load the config from HF,
+            # inject the local tokenizer path, then pass it in.
+            from tada.modules.tada import TadaForCausalLM, TadaConfig
+            logger.info(f"Loading TADA {model_size} model...")
+            config = TadaConfig.from_pretrained(repo)
+            config.tokenizer_name = tokenizer_path
+            self.model = TadaForCausalLM.from_pretrained(
+                repo, config=config, torch_dtype=model_dtype
+            ).to(device)
+            self.model.eval()
+
+        logger.info(f"HumeAI TADA {model_size} loaded successfully on {device}")
+
+    def unload_model(self) -> None:
+        """Unload model and encoder to free memory."""
+        if self.model is not None:
+            del self.model
+            self.model = None
+        if self.encoder is not None:
+            del self.encoder
+            self.encoder = None
+
+        self._device = None
+
+        import torch
+        if torch.cuda.is_available():
+            torch.cuda.empty_cache()
+
+        logger.info("HumeAI TADA unloaded")
+
+    async def create_voice_prompt(
+        self,
+        audio_path: str,
+        reference_text: str,
+        use_cache: bool = True,
+    ) -> Tuple[dict, bool]:
+        """
+        Create voice prompt from reference audio using TADA's encoder.
+
+        TADA's encoder performs forced alignment between audio and text tokens,
+        producing an EncoderOutput with 1:1 token-audio alignment. If no
+        reference_text is provided, the encoder uses built-in ASR (English only).
+
+        We serialize the EncoderOutput to a dict for caching.
+        """
+        await self.load_model(self.model_size)
+
+        cache_key = (
+            "tada_" + get_cache_key(audio_path, reference_text)
+        ) if use_cache else None
+
+        if cache_key:
+            cached = get_cached_voice_prompt(cache_key)
+            if cached is not None and isinstance(cached, dict):
+                return cached, True
+
+        def _encode_sync():
+            import torch
+            import soundfile as sf
+
+            device = self._device
+
+            # Load audio with soundfile (torchaudio 2.10+ requires torchcodec)
+            audio_np, sr = sf.read(str(audio_path), dtype="float32")
+            audio = torch.from_numpy(audio_np).float()
+            if audio.ndim == 1:
+                audio = audio.unsqueeze(0)  # (samples,) -> (1, samples)
+            else:
+                audio = audio.T  # (samples, channels) -> (channels, samples)
+            audio = audio.to(device)
+
+            # Encode with forced alignment
+            text_arg = [reference_text] if reference_text else None
+            prompt = self.encoder(
+                audio, text=text_arg, sample_rate=sr
+            )
+
+            # Serialize EncoderOutput to a dict of CPU tensors for caching
+            prompt_dict = {}
+            for field_name in prompt.__dataclass_fields__:
+                val = getattr(prompt, field_name)
+                if isinstance(val, torch.Tensor):
+                    prompt_dict[field_name] = val.detach().cpu()
+                elif isinstance(val, list):
+                    prompt_dict[field_name] = val
+                elif isinstance(val, (int, float)):
+                    prompt_dict[field_name] = val
+                else:
+                    prompt_dict[field_name] = val
+            return prompt_dict
+
+        encoded = await asyncio.to_thread(_encode_sync)
+
+        if cache_key:
+            cache_voice_prompt(cache_key, encoded)
+
+        return encoded, False
+
+    async def combine_voice_prompts(
+        self,
+        audio_paths: List[str],
+        reference_texts: List[str],
+    ) -> Tuple[np.ndarray, str]:
+        return await _combine_voice_prompts(audio_paths, reference_texts, sample_rate=24000)
+
+    async def generate(
+        self,
+        text: str,
+        voice_prompt: dict,
+        language: str = "en",
+        seed: Optional[int] = None,
+        instruct: Optional[str] = None,
+    ) -> Tuple[np.ndarray, int]:
+        """
+        Generate audio from text using HumeAI TADA.
+
+        Args:
+            text: Text to synthesize
+            voice_prompt: Serialized EncoderOutput dict from create_voice_prompt()
+            language: Language code (en, ar, de, es, fr, it, ja, pl, pt, zh)
+            seed: Random seed for reproducibility
+            instruct: Not supported by TADA (ignored)
+
+        Returns:
+            Tuple of (audio_array, sample_rate=24000)
+        """
+        await self.load_model(self.model_size)
+
+        def _generate_sync():
+            import torch
+            from tada.modules.encoder import EncoderOutput
+
+            if seed is not None:
+                torch.manual_seed(seed)
+                if torch.cuda.is_available():
+                    torch.cuda.manual_seed(seed)
+
+            device = self._device
+
+            # Reconstruct EncoderOutput from the cached dict
+            restored = {}
+            for k, v in voice_prompt.items():
+                if isinstance(v, torch.Tensor):
+                    # Move to device and match model dtype for float tensors
+                    if v.is_floating_point():
+                        model_dtype = next(self.model.parameters()).dtype
+                        restored[k] = v.to(device=device, dtype=model_dtype)
+                    else:
+                        restored[k] = v.to(device=device)
+                else:
+                    restored[k] = v
+
+            prompt = EncoderOutput(**restored)
+
+            # For non-English with the 3B-ML model, we could reload the
+            # encoder with the language-specific aligner. However, the
+            # generation itself is language-agnostic — only the encoder's
+            # aligner changes. Since we encode at create_voice_prompt time,
+            # the language is already baked in. For simplicity, we don't
+            # reload the encoder here.
+
+            logger.info(f"[TADA] Generating ({language}), text length: {len(text)}")
+
+            output = self.model.generate(
+                prompt=prompt,
+                text=text,
+            )
+
+            # output.audio is a list of tensors (one per batch item)
+            if output.audio and output.audio[0] is not None:
+                audio_tensor = output.audio[0]
+                audio = audio_tensor.detach().cpu().numpy().squeeze().astype(np.float32)
+            else:
+                logger.warning("[TADA] Generation produced no audio")
+                audio = np.zeros(24000, dtype=np.float32)
+
+            return audio, 24000
+
+        return await asyncio.to_thread(_generate_sync)

backend/backends/mlx_backend.py

@@ -6,7 +6,6 @@ from typing import Optional, List, Tuple
 import asyncio
 import logging
 import numpy as np
-import os
 from pathlib import Path
 
 logger = logging.getLogger(__name__)
@@ -21,6 +20,7 @@ ensure_original_qwen_config_cached()
 from . import TTSBackend, STTBackend, LANGUAGE_CODE_TO_NAME, WHISPER_HF_REPOS
 from .base import is_model_cached, combine_voice_prompts as _combine_voice_prompts, model_load_progress
 from ..utils.cache import get_cache_key, get_cached_voice_prompt, cache_voice_prompt
+from ..utils.hf_offline_patch import force_offline_if_cached
 
 
 class MLXTTSBackend:
@@ -96,32 +96,13 @@ class MLXTTSBackend:
         model_name = f"qwen-tts-{model_size}"
         is_cached = self._is_model_cached(model_size)
 
-        # Force offline mode when cached to avoid network requests
+        with model_load_progress(model_name, is_cached):
-        original_hf_hub_offline = os.environ.get("HF_HUB_OFFLINE")
+            from mlx_audio.tts import load
-        if is_cached:
-            os.environ["HF_HUB_OFFLINE"] = "1"
-            logger.info("[PATCH] Model %s is cached, forcing HF_HUB_OFFLINE=1 to avoid network requests", model_size)
 
-        try:
+            logger.info("Loading MLX TTS model %s...", model_size)
-            with model_load_progress(model_name, is_cached):
-                from mlx_audio.tts import load
 
-                logger.info("Loading MLX TTS model %s...", model_size)
+            with force_offline_if_cached(is_cached, model_name):
-
+                self.model = load(model_path)
-                try:
-                    self.model = load(model_path)
-                except Exception as load_error:
-                    if is_cached and "offline" in str(load_error).lower():
-                        logger.warning("[PATCH] Offline load failed, trying with network: %s", load_error)
-                        os.environ.pop("HF_HUB_OFFLINE", None)
-                        self.model = load(model_path)
-                    else:
-                        raise
-        finally:
-            if original_hf_hub_offline is not None:
-                os.environ["HF_HUB_OFFLINE"] = original_hf_hub_offline
-            else:
-                os.environ.pop("HF_HUB_OFFLINE", None)
 
         self._current_model_size = model_size
         self.model_size = model_size
@@ -329,7 +310,9 @@ class MLXSTTBackend:
 
             model_name = WHISPER_HF_REPOS.get(model_size, f"openai/whisper-{model_size}")
             logger.info("Loading MLX Whisper model %s...", model_size)
-            self.model = load(model_name)
+
+            with force_offline_if_cached(is_cached, progress_model_name):
+                self.model = load(model_name)
 
         self.model_size = model_size
         logger.info("MLX Whisper model %s loaded successfully", model_size)
@@ -345,18 +328,20 @@ class MLXSTTBackend:
         self,
         audio_path: str,
         language: Optional[str] = None,
+        model_size: Optional[str] = None,
     ) -> str:
         """
         Transcribe audio to text.
 
         Args:
             audio_path: Path to audio file
-            language: Optional language hint (en or zh)
+            language: Optional language hint
+            model_size: Optional model size override
 
         Returns:
             Transcribed text
         """
-        await self.load_model_async(None)
+        await self.load_model_async(model_size)
 
         def _transcribe_sync():
             """Run synchronous transcription in thread pool."""

backend/backends/pytorch_backend.py

@@ -19,6 +19,7 @@ from .base import (
 )
 from ..utils.cache import get_cache_key, get_cached_voice_prompt, cache_voice_prompt
 from ..utils.audio import load_audio
+from ..utils.hf_offline_patch import force_offline_if_cached
 
 
 class PyTorchTTSBackend:
@@ -96,18 +97,19 @@ class PyTorchTTSBackend:
             model_path = self._get_model_path(model_size)
             logger.info("Loading TTS model %s on %s...", model_size, self.device)
 
-            if self.device == "cpu":
+            with force_offline_if_cached(is_cached, model_name):
-                self.model = Qwen3TTSModel.from_pretrained(
+                if self.device == "cpu":
-                    model_path,
+                    self.model = Qwen3TTSModel.from_pretrained(
-                    torch_dtype=torch.float32,
+                        model_path,
-                    low_cpu_mem_usage=False,
+                        torch_dtype=torch.float32,
-                )
+                        low_cpu_mem_usage=False,
-            else:
+                    )
-                self.model = Qwen3TTSModel.from_pretrained(
+                else:
-                    model_path,
+                    self.model = Qwen3TTSModel.from_pretrained(
-                    device_map=self.device,
+                        model_path,
-                    torch_dtype=torch.bfloat16,
+                        device_map=self.device,
-                )
+                        torch_dtype=torch.bfloat16,
+                    )
 
         self._current_model_size = model_size
         self.model_size = model_size
@@ -282,8 +284,9 @@ class PyTorchSTTBackend:
             model_name = WHISPER_HF_REPOS.get(model_size, f"openai/whisper-{model_size}")
             logger.info("Loading Whisper model %s on %s...", model_size, self.device)
 
-            self.processor = WhisperProcessor.from_pretrained(model_name)
+            with force_offline_if_cached(is_cached, progress_model_name):
-            self.model = WhisperForConditionalGeneration.from_pretrained(model_name)
+                self.processor = WhisperProcessor.from_pretrained(model_name)
+                self.model = WhisperForConditionalGeneration.from_pretrained(model_name)
 
         self.model.to(self.device)
         self.model_size = model_size
@@ -306,18 +309,20 @@ class PyTorchSTTBackend:
         self,
         audio_path: str,
         language: Optional[str] = None,
+        model_size: Optional[str] = None,
     ) -> str:
         """
         Transcribe audio to text.
 
         Args:
             audio_path: Path to audio file
-            language: Optional language hint (en or zh)
+            language: Optional language hint
+            model_size: Optional model size override
 
         Returns:
             Transcribed text
         """
-        await self.load_model_async(None)
+        await self.load_model_async(model_size)
 
         def _transcribe_sync():
             """Run synchronous transcription in thread pool."""

backend/build_binary.py

@@ -34,9 +34,15 @@ def build_server(cuda=False):
     binary_name = "voicebox-server-cuda" if cuda else "voicebox-server"
 
     # PyInstaller arguments
+    # CUDA builds use --onedir so we can split the output into two archives:
+    #   1. Server core (~200-400MB) — versioned with the app
+    #   2. CUDA libs (~2GB) — versioned independently (only redownloaded on
+    #      CUDA toolkit / torch major version changes)
+    # CPU builds remain --onefile for simplicity.
+    pack_mode = "--onedir" if cuda else "--onefile"
     args = [
         "server.py",  # Use server.py as entry point instead of main.py
-        "--onefile",
+        pack_mode,
         "--name",
         binary_name,
     ]
@@ -127,7 +133,13 @@ def build_server(cuda=False):
             "uvicorn",
             "--hidden-import",
             "sqlalchemy",
-            "--hidden-import",
+            # librosa uses lazy_loader which generates .pyi stub files at
+            # install time and reads them at runtime to discover submodules.
+            # --hidden-import alone doesn't bundle the stubs, causing
+            # "Cannot load imports from non-existent stub" at runtime.
+            "--collect-all",
+            "lazy_loader",
+            "--collect-all",
             "librosa",
             "--hidden-import",
             "soundfile",
@@ -159,9 +171,9 @@ def build_server(cuda=False):
             "tqdm",
             "--hidden-import",
             "requests",
-            "--collect-submodules",
+            # qwen_tts uses inspect.getsource() at runtime to locate
-            "qwen_tts",
+            # modeling_qwen3_tts.py — needs physical .py source files bundled
-            "--collect-data",
+            "--collect-all",
             "qwen_tts",
             # Fix for pkg_resources and jaraco namespace packages
             "--hidden-import",
@@ -180,6 +192,42 @@ def build_server(cuda=False):
             # needed by LuxTTS for text-to-phoneme conversion
             "--collect-all",
             "piper_phonemize",
+            # HumeAI TADA — speech-language model using Llama + flow matching
+            "--hidden-import",
+            "backend.backends.hume_backend",
+            "--hidden-import",
+            "tada",
+            "--hidden-import",
+            "tada.modules",
+            "--hidden-import",
+            "tada.modules.tada",
+            "--hidden-import",
+            "tada.modules.encoder",
+            "--hidden-import",
+            "tada.modules.decoder",
+            "--hidden-import",
+            "tada.modules.aligner",
+            "--hidden-import",
+            "tada.modules.acoustic_spkr_verf",
+            "--hidden-import",
+            "tada.nn",
+            "--hidden-import",
+            "tada.nn.vibevoice",
+            "--hidden-import",
+            "tada.utils",
+            "--hidden-import",
+            "tada.utils.gray_code",
+            "--hidden-import",
+            "tada.utils.text",
+            # DAC shim — provides dac.nn.layers.Snake1d without the real
+            # descript-audio-codec package (which pulls onnx/tensorboard via
+            # descript-audiotools). The shim is in backend/utils/dac_shim.py.
+            "--hidden-import",
+            "backend.utils.dac_shim",
+            "--hidden-import",
+            "torchaudio",
+            "--collect-submodules",
+            "tada",
         ]
     )
 
@@ -322,7 +370,7 @@ def build_server(cuda=False):
                     "torchvision",
                     "torchaudio",
                     "--index-url",
-                    "https://download.pytorch.org/whl/cu126",
+                    "https://download.pytorch.org/whl/cu128",
                     "--force-reinstall",
                     "-q",
                 ],

backend/models.py

@@ -66,9 +66,9 @@ class GenerationRequest(BaseModel):
     text: str = Field(..., min_length=1, max_length=50000)
     language: str = Field(default="en", pattern="^(zh|en|ja|ko|de|fr|ru|pt|es|it|he|ar|da|el|fi|hi|ms|nl|no|pl|sv|sw|tr)$")
     seed: Optional[int] = Field(None, ge=0)
-    model_size: Optional[str] = Field(default="1.7B", pattern="^(1\\.7B|0\\.6B)$")
+    model_size: Optional[str] = Field(default="1.7B", pattern="^(1\\.7B|0\\.6B|1B|3B)$")
     instruct: Optional[str] = Field(None, max_length=500)
-    engine: Optional[str] = Field(default="qwen", pattern="^(qwen|luxtts|chatterbox|chatterbox_turbo)$")
+    engine: Optional[str] = Field(default="qwen", pattern="^(qwen|luxtts|chatterbox|chatterbox_turbo|tada)$")
     max_chunk_chars: int = Field(
         default=800, ge=100, le=5000, description="Max characters per chunk for long text splitting"
     )
@@ -149,7 +149,8 @@ class HistoryListResponse(BaseModel):
 class TranscriptionRequest(BaseModel):
     """Request model for audio transcription."""
 
-    language: Optional[str] = Field(None, pattern="^(en|zh)$")
+    language: Optional[str] = Field(None, pattern="^(en|zh|ja|ko|de|fr|ru|pt|es|it)$")
+    model: Optional[str] = Field(None, pattern="^(base|small|medium|large|turbo)$")
 
 
 class TranscriptionResponse(BaseModel):

backend/requirements.txt

@@ -8,7 +8,7 @@ sqlalchemy>=2.0.0
 alembic>=1.13.0
 
 # ML models
-torch>=2.1.0
+torch>=2.7.0
 transformers>=4.36.0,<=4.57.6
 accelerate>=0.26.0
 huggingface_hub>=0.20.0
@@ -33,6 +33,13 @@ s3tokenizer
 spacy-pkuseg
 pyloudnorm
 
+# HumeAI TADA sub-dependencies (hume-tada itself is installed
+# --no-deps in the setup script because it pins torch>=2.7,<2.8.
+# descript-audio-codec is NOT installed — it pulls onnx/tensorboard
+# via descript-audiotools.  A lightweight shim in utils/dac_shim.py
+# provides the only class TADA uses: Snake1d.)
+torchaudio
+
 # Audio processing
 librosa>=0.10.0
 soundfile>=0.12.0

backend/routes/generations.py

@@ -1,12 +1,15 @@
 """TTS generation endpoints."""
 
 import asyncio
+import logging
 import uuid
 
 from fastapi import APIRouter, Depends, HTTPException
 from fastapi.responses import StreamingResponse
 from sqlalchemy.orm import Session
 
+logger = logging.getLogger(__name__)
+
 from .. import models
 from ..services import history, profiles, tts
 from ..database import Generation as DBGeneration, VoiceProfile as DBVoiceProfile, get_db
@@ -181,25 +184,28 @@ async def get_generation_status(generation_id: str, db: Session = Depends(get_db
     import json
 
     async def event_stream():
-        while True:
+        try:
-            db.expire_all()
+            while True:
-            gen = db.query(DBGeneration).filter_by(id=generation_id).first()
+                db.expire_all()
-            if not gen:
+                gen = db.query(DBGeneration).filter_by(id=generation_id).first()
-                yield f"data: {json.dumps({'status': 'not_found', 'id': generation_id})}\n\n"
+                if not gen:
-                return
+                    yield f"data: {json.dumps({'status': 'not_found', 'id': generation_id})}\n\n"
+                    return
 
-            payload = {
+                payload = {
-                "id": gen.id,
+                    "id": gen.id,
-                "status": gen.status or "completed",
+                    "status": gen.status or "completed",
-                "duration": gen.duration,
+                    "duration": gen.duration,
-                "error": gen.error,
+                    "error": gen.error,
-            }
+                }
-            yield f"data: {json.dumps(payload)}\n\n"
+                yield f"data: {json.dumps(payload)}\n\n"
 
-            if (gen.status or "completed") in ("completed", "failed"):
+                if (gen.status or "completed") in ("completed", "failed"):
-                return
+                    return
 
-            await asyncio.sleep(1)
+                await asyncio.sleep(1)
+        except (BrokenPipeError, ConnectionResetError, asyncio.CancelledError):
+            logger.debug("SSE client disconnected for generation %s", generation_id)
 
     return StreamingResponse(
         event_stream(),
@@ -265,9 +271,12 @@ async def stream_speech(
     wav_bytes = tts.audio_to_wav_bytes(audio, sample_rate)
 
     async def _wav_stream():
-        chunk_size = 64 * 1024
+        try:
-        for i in range(0, len(wav_bytes), chunk_size):
+            chunk_size = 64 * 1024
-            yield wav_bytes[i : i + chunk_size]
+            for i in range(0, len(wav_bytes), chunk_size):
+                yield wav_bytes[i : i + chunk_size]
+        except (BrokenPipeError, ConnectionResetError, asyncio.CancelledError):
+            logger.debug("Client disconnected during audio stream")
 
     return StreamingResponse(
         _wav_stream(),

backend/routes/health.py

@@ -3,9 +3,11 @@
 import asyncio
 import os
 import signal
+from pathlib import Path
 
 import torch
 from fastapi import APIRouter, Depends
+from fastapi.responses import FileResponse
 from sqlalchemy.orm import Session
 
 from .. import config, models
@@ -15,12 +17,18 @@ from ..utils.platform_detect import get_backend_type
 
 router = APIRouter()
 
+# Frontend build directory — present in Docker, absent in dev/API-only mode
+_frontend_dir = Path(__file__).resolve().parent.parent.parent / "frontend"
+
 
 @router.get("/")
 async def root():
-    """Root endpoint."""
+    """Root endpoint — serves SPA index.html in Docker, JSON otherwise."""
     from .. import __version__
 
+    index = _frontend_dir / "index.html"
+    if index.is_file():
+        return FileResponse(index, media_type="text/html")
     return {"message": "voicebox API", "version": __version__}
 
 
@@ -199,7 +207,7 @@ async def filesystem_health():
 
         checks.append(
             models.DirectoryCheck(
-                path=str(dir_path),
+                path=str(dir_path.resolve()),
                 exists=exists,
                 writable=writable,
                 error=error,

backend/routes/profiles.py

@@ -102,6 +102,10 @@ async def delete_profile(
     return {"message": "Profile deleted successfully"}
 
 
+SAMPLE_MAX_FILE_SIZE = 50 * 1024 * 1024  # 50 MB
+SAMPLE_UPLOAD_CHUNK_SIZE = 1024 * 1024  # 1 MB
+
+
 @router.post("/profiles/{profile_id}/samples", response_model=models.ProfileSampleResponse)
 async def add_profile_sample(
     profile_id: str,
@@ -115,8 +119,16 @@ async def add_profile_sample(
     file_suffix = _uploaded_ext if _uploaded_ext in _allowed_audio_exts else ".wav"
 
     with tempfile.NamedTemporaryFile(suffix=file_suffix, delete=False) as tmp:
-        content = await file.read()
+        total_size = 0
-        tmp.write(content)
+        while chunk := await file.read(SAMPLE_UPLOAD_CHUNK_SIZE):
+            total_size += len(chunk)
+            if total_size > SAMPLE_MAX_FILE_SIZE:
+                Path(tmp.name).unlink(missing_ok=True)
+                raise HTTPException(
+                    status_code=413,
+                    detail=f"File too large (max {SAMPLE_MAX_FILE_SIZE // (1024 * 1024)} MB)",
+                )
+            tmp.write(chunk)
         tmp_path = tmp.name
 
     try:

backend/routes/transcription.py

@@ -20,6 +20,7 @@ UPLOAD_CHUNK_SIZE = 1024 * 1024  # 1MB
 async def transcribe_audio(
     file: UploadFile = File(...),
     language: str | None = Form(None),
+    model: str | None = Form(None),
 ):
     """Transcribe audio file to text."""
     with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as tmp:
@@ -29,14 +30,23 @@ async def transcribe_audio(
 
     try:
         from ..utils.audio import load_audio
+        from ..backends import WHISPER_HF_REPOS
 
         audio, sr = await asyncio.to_thread(load_audio, tmp_path)
         duration = len(audio) / sr
 
         whisper_model = transcribe.get_whisper_model()
-        model_size = whisper_model.model_size
+        model_size = model if model else whisper_model.model_size
 
-        if not whisper_model.is_loaded() and not whisper_model._is_model_cached(model_size):
+        valid_sizes = list(WHISPER_HF_REPOS.keys())
+        if model_size not in valid_sizes:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Invalid model size '{model_size}'. Must be one of: {', '.join(valid_sizes)}",
+            )
+
+        already_loaded = whisper_model.is_loaded() and whisper_model.model_size == model_size
+        if not already_loaded and not whisper_model._is_model_cached(model_size):
             progress_model_name = f"whisper-{model_size}"
             task_manager = get_task_manager()
 
@@ -59,7 +69,7 @@ async def transcribe_audio(
                 },
             )
 
-        text = await whisper_model.transcribe(tmp_path, language)
+        text = await whisper_model.transcribe(tmp_path, language, model_size)
 
         return models.TranscriptionResponse(
             text=text,

backend/services/cuda.py

@@ -1,16 +1,22 @@
 """
-CUDA backend binary download, assembly, and verification.
+CUDA backend download, assembly, and verification.
 
-Downloads split parts of the CUDA-enabled voicebox-server binary from
+Downloads two archives from GitHub Releases:
-GitHub Releases, reassembles them, verifies integrity via SHA-256,
+  1. Server core (voicebox-server-cuda.tar.gz) — the exe + non-NVIDIA deps,
-and places the binary in the app's data directory for use on next
+     versioned with the app.
-backend restart.
+  2. CUDA libs (cuda-libs-{version}.tar.gz) — NVIDIA runtime libraries,
+     versioned independently (only redownloaded on CUDA toolkit bump).
+
+Both archives are extracted into {data_dir}/backends/cuda/ which forms the
+complete PyInstaller --onedir directory structure that torch expects.
 """
 
 import hashlib
+import json
 import logging
 import os
 import sys
+import tarfile
 from pathlib import Path
 from typing import Optional
 
@@ -24,6 +30,10 @@ GITHUB_RELEASES_URL = "https://github.com/jamiepine/voicebox/releases/download"
 
 PROGRESS_KEY = "cuda-backend"
 
+# The current expected CUDA libs version.  Bump this when we change the
+# CUDA toolkit version or torch's CUDA dependency changes (e.g. cu126 -> cu128).
+CUDA_LIBS_VERSION = "cu128-v1"
+
 
 def get_backends_dir() -> Path:
     """Directory where downloaded backend binaries are stored."""
@@ -32,21 +42,46 @@ def get_backends_dir() -> Path:
     return d
 
 
-def get_cuda_binary_name() -> str:
+def get_cuda_dir() -> Path:
-    """Platform-specific CUDA binary filename."""
+    """Directory where the CUDA backend (onedir) is extracted."""
+    d = get_backends_dir() / "cuda"
+    d.mkdir(parents=True, exist_ok=True)
+    return d
+
+
+def get_cuda_exe_name() -> str:
+    """Platform-specific CUDA executable filename."""
     if sys.platform == "win32":
         return "voicebox-server-cuda.exe"
     return "voicebox-server-cuda"
 
 
 def get_cuda_binary_path() -> Optional[Path]:
-    """Return path to CUDA binary if it exists."""
+    """Return path to the CUDA executable if it exists inside the onedir."""
-    p = get_backends_dir() / get_cuda_binary_name()
+    p = get_cuda_dir() / get_cuda_exe_name()
     if p.exists():
         return p
     return None
 
 
+def get_cuda_libs_manifest_path() -> Path:
+    """Path to the cuda-libs.json manifest inside the CUDA dir."""
+    return get_cuda_dir() / "cuda-libs.json"
+
+
+def get_installed_cuda_libs_version() -> Optional[str]:
+    """Read the installed CUDA libs version from cuda-libs.json, or None."""
+    manifest_path = get_cuda_libs_manifest_path()
+    if not manifest_path.exists():
+        return None
+    try:
+        data = json.loads(manifest_path.read_text())
+        return data.get("version")
+    except Exception as e:
+        logger.warning(f"Could not read cuda-libs.json: {e}")
+        return None
+
+
 def is_cuda_active() -> bool:
     """Check if the current process is the CUDA binary.
 
@@ -60,25 +95,151 @@ def get_cuda_status() -> dict:
     progress_manager = get_progress_manager()
     cuda_path = get_cuda_binary_path()
     progress = progress_manager.get_progress(PROGRESS_KEY)
+    cuda_libs_version = get_installed_cuda_libs_version()
 
     return {
         "available": cuda_path is not None,
         "active": is_cuda_active(),
         "binary_path": str(cuda_path) if cuda_path else None,
+        "cuda_libs_version": cuda_libs_version,
         "downloading": progress is not None and progress.get("status") == "downloading",
         "download_progress": progress,
     }
 
 
-async def download_cuda_binary(version: Optional[str] = None):
+def _needs_server_download(version: Optional[str] = None) -> bool:
-    """Download the CUDA backend binary from GitHub Releases.
+    """Check if the server core archive needs to be (re)downloaded."""
+    cuda_path = get_cuda_binary_path()
+    if not cuda_path:
+        return True
+    # Check if the binary version matches the expected app version
+    installed = get_cuda_binary_version()
+    expected = version or __version__
+    if expected.startswith("v"):
+        expected = expected[1:]
+    return installed != expected
 
-    Downloads split parts listed in a manifest file, concatenates them,
+
-    and verifies the SHA-256 checksum for integrity. Atomic write
+def _needs_cuda_libs_download() -> bool:
-    (temp file -> rename).
+    """Check if the CUDA libs archive needs to be (re)downloaded."""
+    installed = get_installed_cuda_libs_version()
+    if installed is None:
+        return True
+    return installed != CUDA_LIBS_VERSION
+
+
+async def _download_and_extract_archive(
+    client,
+    url: str,
+    sha256_url: Optional[str],
+    dest_dir: Path,
+    label: str,
+    progress_offset: int,
+    total_size: int,
+):
+    """Download a .tar.gz archive and extract it into dest_dir.
 
     Args:
-        version: Version tag (e.g. "v0.2.0"). Defaults to current app version.
+        client: httpx.AsyncClient
+        url: URL of the .tar.gz archive
+        sha256_url: URL of the .sha256 checksum file (optional)
+        dest_dir: Directory to extract into
+        label: Human-readable label for progress updates
+        progress_offset: Byte offset for progress reporting (when downloading
+            multiple archives sequentially)
+        total_size: Total bytes across all downloads (for progress bar)
+    """
+    progress = get_progress_manager()
+    temp_path = dest_dir / f".download-{label.replace(' ', '-')}.tmp"
+
+    # Clean up leftover partial download
+    if temp_path.exists():
+        temp_path.unlink()
+
+    # Fetch expected checksum (fail-fast: never extract an unverified archive)
+    expected_sha = None
+    if sha256_url:
+        try:
+            sha_resp = await client.get(sha256_url)
+            sha_resp.raise_for_status()
+            expected_sha = sha_resp.text.strip().split()[0]
+            logger.info(f"{label}: expected SHA-256: {expected_sha[:16]}...")
+        except Exception as e:
+            raise RuntimeError(f"{label}: failed to fetch checksum from {sha256_url}") from e
+
+    # Stream download, verify, and extract — always clean up temp file
+    downloaded = 0
+    try:
+        async with client.stream("GET", url) as response:
+            response.raise_for_status()
+            with open(temp_path, "wb") as f:
+                async for chunk in response.aiter_bytes(chunk_size=1024 * 1024):
+                    f.write(chunk)
+                    downloaded += len(chunk)
+                    progress.update_progress(
+                        PROGRESS_KEY,
+                        current=progress_offset + downloaded,
+                        total=total_size,
+                        filename=f"Downloading {label}",
+                        status="downloading",
+                    )
+
+        # Verify integrity
+        if expected_sha:
+            progress.update_progress(
+                PROGRESS_KEY,
+                current=progress_offset + downloaded,
+                total=total_size,
+                filename=f"Verifying {label}...",
+                status="downloading",
+            )
+            sha256 = hashlib.sha256()
+            with open(temp_path, "rb") as f:
+                while True:
+                    data = f.read(1024 * 1024)
+                    if not data:
+                        break
+                    sha256.update(data)
+            actual = sha256.hexdigest()
+            if actual != expected_sha:
+                raise ValueError(
+                    f"{label} integrity check failed: expected {expected_sha[:16]}..., got {actual[:16]}..."
+                )
+            logger.info(f"{label}: integrity verified")
+
+        # Extract (use data filter for path traversal protection on Python 3.12+)
+        progress.update_progress(
+            PROGRESS_KEY,
+            current=progress_offset + downloaded,
+            total=total_size,
+            filename=f"Extracting {label}...",
+            status="downloading",
+        )
+        with tarfile.open(temp_path, "r:gz") as tar:
+            if sys.version_info >= (3, 12):
+                tar.extractall(path=dest_dir, filter="data")
+            else:
+                tar.extractall(path=dest_dir)
+
+        logger.info(f"{label}: extracted to {dest_dir}")
+    finally:
+        if temp_path.exists():
+            temp_path.unlink()
+    return downloaded
+
+
+async def download_cuda_binary(version: Optional[str] = None):
+    """Download the CUDA backend (server core + CUDA libs if needed).
+
+    Downloads both archives from GitHub Releases, extracts them into
+    {data_dir}/backends/cuda/, and writes the cuda-libs.json manifest.
+
+    Only downloads what's needed:
+    - Server core: always redownloaded (versioned with app)
+    - CUDA libs: only if missing or version mismatch
+
+    Args:
+        version: Version tag (e.g. "v0.3.0"). Defaults to current app version.
     """
     import httpx
 
@@ -86,114 +247,91 @@ async def download_cuda_binary(version: Optional[str] = None):
         version = f"v{__version__}"
 
     progress = get_progress_manager()
-    binary_name = get_cuda_binary_name()
+    cuda_dir = get_cuda_dir()
-    dest_dir = get_backends_dir()
-    final_path = dest_dir / binary_name
-    temp_path = dest_dir / f"{binary_name}.download"
 
-    # Clean up any leftover partial download
+    need_server = _needs_server_download(version)
-    if temp_path.exists():
+    need_libs = _needs_cuda_libs_download()
-        temp_path.unlink()
 
-    logger.info(f"Starting CUDA backend download for {version}")
+    if not need_server and not need_libs:
+        logger.info("CUDA backend is up to date, nothing to download")
+        return
+
+    logger.info(
+        f"Starting CUDA backend download for {version} "
+        f"(server={'yes' if need_server else 'cached'}, "
+        f"libs={'yes' if need_libs else 'cached'})"
+    )
     progress.update_progress(
-        PROGRESS_KEY, current=0, total=0,
+        PROGRESS_KEY,
-        filename="Fetching manifest...", status="downloading",
+        current=0,
+        total=0,
+        filename="Preparing download...",
+        status="downloading",
     )
 
     base_url = f"{GITHUB_RELEASES_URL}/{version}"
-    stem = Path(binary_name).stem  # voicebox-server-cuda
+    server_archive = "voicebox-server-cuda.tar.gz"
+    libs_archive = f"cuda-libs-{CUDA_LIBS_VERSION}.tar.gz"
 
     try:
         async with httpx.AsyncClient(follow_redirects=True, timeout=30.0) as client:
-            # Fetch the manifest (list of split part filenames)
+            # Estimate total download size
-            manifest_url = f"{base_url}/{stem}.manifest"
-            manifest_resp = await client.get(manifest_url)
-            manifest_resp.raise_for_status()
-            parts = [p.strip() for p in manifest_resp.text.strip().splitlines() if p.strip()]
-
-            if not parts:
-                raise ValueError("Empty manifest — no split parts found")
-
-            logger.info(f"Found {len(parts)} split parts to download")
-
-            # Fetch expected checksum (optional — for integrity verification)
-            expected_sha = None
-            try:
-                sha_url = f"{base_url}/{stem}.sha256"
-                sha_resp = await client.get(sha_url)
-                if sha_resp.status_code == 200:
-                    # Format: "sha256hex  filename\n"
-                    expected_sha = sha_resp.text.strip().split()[0]
-                    logger.info(f"Expected SHA-256: {expected_sha[:16]}...")
-            except Exception as e:
-                logger.warning(f"Could not fetch checksum file — skipping verification: {e}")
-
-            # Get total size across all parts by issuing HEAD requests
             total_size = 0
-            for part_name in parts:
+            if need_server:
                 try:
-                    head_resp = await client.head(f"{base_url}/{part_name}")
+                    head = await client.head(f"{base_url}/{server_archive}")
-                    content_length = int(head_resp.headers.get("content-length", 0))
+                    total_size += int(head.headers.get("content-length", 0))
-                    total_size += content_length
                 except Exception:
                     pass
+            if need_libs:
+                try:
+                    head = await client.head(f"{base_url}/{libs_archive}")
+                    total_size += int(head.headers.get("content-length", 0))
+                except Exception:
+                    pass
+
             logger.info(f"Total download size: {total_size / 1024 / 1024:.1f} MB")
 
-            # Download and concatenate parts
+            offset = 0
-            total_downloaded = 0
-            with open(temp_path, "wb") as f:
-                for i, part_name in enumerate(parts):
-                    part_url = f"{base_url}/{part_name}"
-                    logger.info(f"Downloading part {i + 1}/{len(parts)}: {part_name}")
 
-                    async with client.stream("GET", part_url) as response:
+            # Download server core
-                        response.raise_for_status()
+            if need_server:
-                        async for chunk in response.aiter_bytes(chunk_size=1024 * 1024):
+                server_downloaded = await _download_and_extract_archive(
-                            f.write(chunk)
+                    client,
-                            total_downloaded += len(chunk)
+                    url=f"{base_url}/{server_archive}",
-                            progress.update_progress(
+                    sha256_url=f"{base_url}/{server_archive}.sha256",
-                                PROGRESS_KEY, current=total_downloaded, total=total_size,
+                    dest_dir=cuda_dir,
-                                filename=f"Downloading CUDA backend ({i + 1}/{len(parts)})",
+                    label="CUDA server",
-                                status="downloading",
+                    progress_offset=offset,
-                            )
+                    total_size=total_size,
-
-        # Verify integrity if checksum was available
-        if expected_sha:
-            progress.update_progress(
-                PROGRESS_KEY, current=total_downloaded, total=total_downloaded,
-                filename="Verifying integrity...", status="downloading",
-            )
-            sha256 = hashlib.sha256()
-            with open(temp_path, "rb") as f:
-                while True:
-                    chunk = f.read(1024 * 1024)
-                    if not chunk:
-                        break
-                    sha256.update(chunk)
-
-            actual = sha256.hexdigest()
-            if actual != expected_sha:
-                raise ValueError(
-                    f"Integrity check failed: expected {expected_sha[:16]}..., "
-                    f"got {actual[:16]}..."
                 )
-            logger.info(f"Integrity verified: {actual[:16]}...")
+                offset += server_downloaded
 
-        # Atomic move into place (replace handles existing target on all platforms)
+                # Make executable on Unix
-        temp_path.replace(final_path)
+                exe_path = cuda_dir / get_cuda_exe_name()
+                if sys.platform != "win32" and exe_path.exists():
+                    exe_path.chmod(0o755)
 
-        # Make executable on Unix
+            # Download CUDA libs
-        if sys.platform != "win32":
+            if need_libs:
-            final_path.chmod(0o755)
+                await _download_and_extract_archive(
+                    client,
+                    url=f"{base_url}/{libs_archive}",
+                    sha256_url=f"{base_url}/{libs_archive}.sha256",
+                    dest_dir=cuda_dir,
+                    label="CUDA libraries",
+                    progress_offset=offset,
+                    total_size=total_size,
+                )
 
-        logger.info(f"CUDA backend downloaded to {final_path}")
+                # Write local cuda-libs.json manifest
+                manifest = {"version": CUDA_LIBS_VERSION}
+                get_cuda_libs_manifest_path().write_text(json.dumps(manifest, indent=2) + "\n")
+
+        logger.info(f"CUDA backend ready at {cuda_dir}")
         progress.mark_complete(PROGRESS_KEY)
 
     except Exception as e:
-        # Clean up on failure
-        if temp_path.exists():
-            temp_path.unlink()
         logger.error(f"CUDA backend download failed: {e}")
         progress.mark_error(PROGRESS_KEY, str(e))
         raise
@@ -202,15 +340,19 @@ async def download_cuda_binary(version: Optional[str] = None):
 def get_cuda_binary_version() -> Optional[str]:
     """Get the version of the installed CUDA binary, or None if not installed."""
     import subprocess
+
     cuda_path = get_cuda_binary_path()
     if not cuda_path:
         return None
     try:
         result = subprocess.run(
             [str(cuda_path), "--version"],
-            capture_output=True, text=True, timeout=30,
+            capture_output=True,
+            text=True,
+            timeout=30,
+            cwd=str(cuda_path.parent),  # Run from the onedir directory
         )
-        # Output format: "voicebox-server 0.2.0"
+        # Output format: "voicebox-server 0.3.0"
         for line in result.stdout.strip().splitlines():
             if "voicebox-server" in line:
                 return line.split()[-1]
@@ -222,26 +364,29 @@ def get_cuda_binary_version() -> Optional[str]:
 async def check_and_update_cuda_binary():
     """Check if the CUDA binary is outdated and auto-download if so.
 
-    Called on server startup. If a CUDA binary exists but its version
+    Called on server startup. Checks both server version and CUDA libs
-    doesn't match the current app version, triggers a background download
+    version. Downloads only what's needed.
-    of the updated CUDA binary. The download progress is visible to the
-    frontend via the existing SSE progress endpoint.
     """
     cuda_path = get_cuda_binary_path()
     if not cuda_path:
         return  # No CUDA binary installed, nothing to update
 
-    cuda_version = get_cuda_binary_version()
+    need_server = _needs_server_download()
-    current_version = __version__
+    need_libs = _needs_cuda_libs_download()
 
-    if cuda_version == current_version:
+    if not need_server and not need_libs:
-        logger.info(f"CUDA binary is up to date (v{current_version})")
+        logger.info(f"CUDA binary is up to date (server=v{__version__}, libs={get_installed_cuda_libs_version()})")
         return
 
-    logger.info(
+    reasons = []
-        f"CUDA binary version mismatch: binary=v{cuda_version}, app=v{current_version}. "
+    if need_server:
-        f"Auto-downloading updated CUDA backend..."
+        cuda_version = get_cuda_binary_version()
-    )
+        reasons.append(f"server v{cuda_version} != v{__version__}")
+    if need_libs:
+        installed_libs = get_installed_cuda_libs_version()
+        reasons.append(f"libs {installed_libs} != {CUDA_LIBS_VERSION}")
+
+    logger.info(f"CUDA backend needs update ({', '.join(reasons)}). Auto-downloading...")
 
     try:
         await download_cuda_binary()
@@ -250,10 +395,12 @@ async def check_and_update_cuda_binary():
 
 
 async def delete_cuda_binary() -> bool:
-    """Delete the downloaded CUDA binary. Returns True if deleted."""
+    """Delete the downloaded CUDA backend directory. Returns True if deleted."""
-    path = get_cuda_binary_path()
+    import shutil
-    if path and path.exists():
+
-        path.unlink()
+    cuda_dir = get_cuda_dir()
-        logger.info(f"Deleted CUDA binary: {path}")
+    if cuda_dir.exists() and any(cuda_dir.iterdir()):
+        shutil.rmtree(cuda_dir)
+        logger.info(f"Deleted CUDA backend directory: {cuda_dir}")
         return True
     return False

backend/services/profiles.py

@@ -22,7 +22,7 @@ from ..database import (
     Generation as DBGeneration,
 )
 from ..models import EffectConfig
-from ..utils.audio import validate_reference_audio, load_audio, save_audio
+from ..utils.audio import validate_reference_audio, validate_and_load_reference_audio, load_audio, save_audio
 from ..utils.images import validate_image, process_avatar
 from ..utils.cache import _get_cache_dir, clear_profile_cache
 from .tts import get_tts_model
@@ -117,11 +117,16 @@ async def add_profile_sample(
     Returns:
         Created sample
     """
+    import asyncio
+
     profile = db.query(DBVoiceProfile).filter_by(id=profile_id).first()
     if not profile:
         raise ValueError(f"Profile {profile_id} not found")
 
-    is_valid, error_msg = validate_reference_audio(audio_path)
+    # Validate and load audio in a single pass, off the event loop
+    is_valid, error_msg, audio, sr = await asyncio.to_thread(
+        validate_and_load_reference_audio, audio_path
+    )
     if not is_valid:
         raise ValueError(f"Invalid reference audio: {error_msg}")
 
@@ -130,8 +135,7 @@ async def add_profile_sample(
     profile_dir.mkdir(parents=True, exist_ok=True)
 
     dest_path = profile_dir / f"{sample_id}.wav"
-    audio, sr = load_audio(audio_path)
+    await asyncio.to_thread(save_audio, audio, str(dest_path), sr)
-    save_audio(audio, str(dest_path), sr)
 
     db_sample = DBProfileSample(
         id=sample_id,

backend/utils/audio.py

@@ -217,22 +217,40 @@ def validate_reference_audio(
     Returns:
         Tuple of (is_valid, error_message)
     """
+    result = validate_and_load_reference_audio(
+        audio_path, min_duration, max_duration, min_rms
+    )
+    return (result[0], result[1])
+
+
+def validate_and_load_reference_audio(
+    audio_path: str,
+    min_duration: float = 2.0,
+    max_duration: float = 30.0,
+    min_rms: float = 0.01,
+) -> Tuple[bool, Optional[str], Optional[np.ndarray], Optional[int]]:
+    """
+    Validate and load reference audio in a single pass.
+    
+    Returns:
+        Tuple of (is_valid, error_message, audio_array, sample_rate)
+    """
     try:
         audio, sr = load_audio(audio_path)
         duration = len(audio) / sr
         
         if duration < min_duration:
-            return False, f"Audio too short (minimum {min_duration} seconds)"
+            return False, f"Audio too short (minimum {min_duration} seconds)", None, None
         if duration > max_duration:
-            return False, f"Audio too long (maximum {max_duration} seconds)"
+            return False, f"Audio too long (maximum {max_duration} seconds)", None, None
         
         rms = np.sqrt(np.mean(audio**2))
         if rms < min_rms:
-            return False, "Audio is too quiet or silent"
+            return False, "Audio is too quiet or silent", None, None
         
         if np.abs(audio).max() > 0.99:
-            return False, "Audio is clipping (reduce input gain)"
+            return False, "Audio is clipping (reduce input gain)", None, None
         
-        return True, None
+        return True, None, audio, sr
     except Exception as e:
-        return False, f"Error validating audio: {str(e)}"
+        return False, f"Error validating audio: {str(e)}", None, None

backend/utils/dac_shim.py

@@ -0,0 +1,95 @@
+"""
+Minimal shim for descript-audio-codec (DAC).
+
+TADA only imports Snake1d from dac.nn.layers and dac.model.dac.
+The real DAC package pulls in descript-audiotools which depends on
+onnx, tensorboard, protobuf, matplotlib, pystoi, etc. — none of
+which are needed for TADA's runtime use of Snake1d.
+
+This shim provides the exact Snake1d implementation (MIT-licensed,
+from https://github.com/descriptinc/descript-audio-codec) so we can
+avoid the entire audiotools dependency chain.
+
+If the real DAC package is installed, this module is never used —
+Python's import system will find the site-packages version first.
+Install this shim only when descript-audio-codec is NOT installed.
+"""
+
+import sys
+import types
+
+import torch
+import torch.nn as nn
+
+
+# ── Snake activation (from dac/nn/layers.py) ────────────────────────
+
+# NOTE: The original DAC code uses @torch.jit.script here for a 1.4x
+# speedup.  We omit it because TorchScript calls inspect.getsource()
+# which fails inside a PyInstaller frozen binary (no .py source files).
+def snake(x: torch.Tensor, alpha: torch.Tensor) -> torch.Tensor:
+    shape = x.shape
+    x = x.reshape(shape[0], shape[1], -1)
+    x = x + (alpha + 1e-9).reciprocal() * torch.sin(alpha * x).pow(2)
+    x = x.reshape(shape)
+    return x
+
+
+class Snake1d(nn.Module):
+    def __init__(self, channels: int):
+        super().__init__()
+        self.alpha = nn.Parameter(torch.ones(1, channels, 1))
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return snake(x, self.alpha)
+
+
+# ── Register as dac.nn.layers and dac.model.dac ─────────────────────
+
+def install_dac_shim() -> None:
+    """Register fake dac package modules in sys.modules.
+
+    Only installs the shim if 'dac' is not already importable
+    (i.e. the real descript-audio-codec is not installed).
+    """
+    try:
+        import dac  # noqa: F401  — real package exists, do nothing
+        return
+    except ImportError:
+        pass
+
+    # Create the module tree: dac -> dac.nn -> dac.nn.layers
+    #                              -> dac.model -> dac.model.dac
+    dac_pkg = types.ModuleType("dac")
+    dac_pkg.__path__ = []  # make it a package
+    dac_pkg.__package__ = "dac"
+
+    dac_nn = types.ModuleType("dac.nn")
+    dac_nn.__path__ = []
+    dac_nn.__package__ = "dac.nn"
+
+    dac_nn_layers = types.ModuleType("dac.nn.layers")
+    dac_nn_layers.__package__ = "dac.nn"
+    dac_nn_layers.Snake1d = Snake1d
+    dac_nn_layers.snake = snake
+
+    dac_model = types.ModuleType("dac.model")
+    dac_model.__path__ = []
+    dac_model.__package__ = "dac.model"
+
+    dac_model_dac = types.ModuleType("dac.model.dac")
+    dac_model_dac.__package__ = "dac.model"
+    dac_model_dac.Snake1d = Snake1d
+
+    # Wire up submodules
+    dac_pkg.nn = dac_nn
+    dac_pkg.model = dac_model
+    dac_nn.layers = dac_nn_layers
+    dac_model.dac = dac_model_dac
+
+    # Register in sys.modules
+    sys.modules["dac"] = dac_pkg
+    sys.modules["dac.nn"] = dac_nn
+    sys.modules["dac.nn.layers"] = dac_nn_layers
+    sys.modules["dac.model"] = dac_model
+    sys.modules["dac.model.dac"] = dac_model_dac

backend/utils/hf_offline_patch.py

@@ -1,17 +1,64 @@
 """Monkey-patch huggingface_hub to force offline mode with cached models.
 
-Prevents mlx_audio from making network requests when models are already
+Prevents mlx_audio / transformers from making network requests when models
-downloaded. Must be imported BEFORE mlx_audio.
+are already downloaded. Must be imported BEFORE mlx_audio.
 """
 
 import logging
 import os
+from contextlib import contextmanager
 from pathlib import Path
 from typing import Optional, Union
 
 logger = logging.getLogger(__name__)
 
 
+@contextmanager
+def force_offline_if_cached(is_cached: bool, model_label: str = ""):
+    """Context manager that sets ``HF_HUB_OFFLINE=1`` while loading a cached model.
+
+    If *is_cached* is ``False`` the block runs normally (network allowed).
+    If the offline load raises an error containing "offline" we automatically
+    retry with network access so a partially-cached model still works.
+
+    Args:
+        is_cached: Whether the model weights are already on disk.
+        model_label: Human-readable name used in log messages.
+    """
+    if not is_cached:
+        yield
+        return
+
+    original_value = os.environ.get("HF_HUB_OFFLINE")
+    os.environ["HF_HUB_OFFLINE"] = "1"
+    logger.info(
+        "[offline-guard] %s is cached — forcing HF_HUB_OFFLINE=1",
+        model_label or "model",
+    )
+
+    try:
+        yield
+    except Exception as exc:
+        if "offline" in str(exc).lower():
+            logger.warning(
+                "[offline-guard] Offline load failed for %s, retrying with network: %s",
+                model_label or "model",
+                exc,
+            )
+            # Restore original env and retry — caller must wrap the load
+            # inside force_offline_if_cached so retrying here isn't possible.
+            # Instead, propagate a flag via the exception so the caller can
+            # decide.  For simplicity we just let it fall through to the
+            # finally block and re-raise.
+            raise
+        raise
+    finally:
+        if original_value is not None:
+            os.environ["HF_HUB_OFFLINE"] = original_value
+        else:
+            os.environ.pop("HF_HUB_OFFLINE", None)
+
+
 def patch_huggingface_hub_offline():
     """Monkey-patch huggingface_hub to force offline mode."""
     try:

backend/utils/progress.py

@@ -246,6 +246,8 @@ class ProgressManager:
                     # Send heartbeat
                     yield ": heartbeat\n\n"
                     continue
+        except (BrokenPipeError, ConnectionResetError, asyncio.CancelledError):
+            logger.debug(f"SSE client disconnected from {model_name}")
         finally:
             # Remove from listeners
             if model_name in self._listeners:

backend/voicebox-server.spec

@@ -6,7 +6,7 @@ from PyInstaller.utils.hooks import copy_metadata
 
 datas = []
 binaries = []
-hiddenimports = ['backend', 'backend.main', 'backend.config', 'backend.database', 'backend.models', 'backend.profiles', 'backend.history', 'backend.tts', 'backend.transcribe', 'backend.platform_detect', 'backend.backends', 'backend.backends.pytorch_backend', 'backend.utils.audio', 'backend.utils.cache', 'backend.utils.progress', 'backend.utils.hf_progress', 'backend.utils.validation', 'backend.cuda_download', 'backend.effects', 'backend.utils.effects', 'backend.versions', 'pedalboard', 'chatterbox', 'chatterbox.tts_turbo', 'chatterbox.mtl_tts', 'backend.backends.chatterbox_backend', 'backend.backends.chatterbox_turbo_backend', 'backend.backends.luxtts_backend', 'zipvoice', 'zipvoice.luxvoice', 'torch', 'transformers', 'fastapi', 'uvicorn', 'sqlalchemy', 'librosa', 'soundfile', 'qwen_tts', 'qwen_tts.inference', 'qwen_tts.inference.qwen3_tts_model', 'qwen_tts.inference.qwen3_tts_tokenizer', 'qwen_tts.core', 'qwen_tts.cli', 'requests', 'pkg_resources.extern', 'backend.backends.mlx_backend', 'mlx', 'mlx.core', 'mlx.nn', 'mlx_audio', 'mlx_audio.tts', 'mlx_audio.stt']
+hiddenimports = ['backend', 'backend.main', 'backend.config', 'backend.database', 'backend.models', 'backend.services.profiles', 'backend.services.history', 'backend.services.tts', 'backend.services.transcribe', 'backend.utils.platform_detect', 'backend.backends', 'backend.backends.pytorch_backend', 'backend.utils.audio', 'backend.utils.cache', 'backend.utils.progress', 'backend.utils.hf_progress', 'backend.services.cuda', 'backend.services.effects', 'backend.utils.effects', 'backend.services.versions', 'pedalboard', 'chatterbox', 'chatterbox.tts_turbo', 'chatterbox.mtl_tts', 'backend.backends.chatterbox_backend', 'backend.backends.chatterbox_turbo_backend', 'backend.backends.luxtts_backend', 'zipvoice', 'zipvoice.luxvoice', 'torch', 'transformers', 'fastapi', 'uvicorn', 'sqlalchemy', 'soundfile', 'qwen_tts', 'qwen_tts.inference', 'qwen_tts.inference.qwen3_tts_model', 'qwen_tts.inference.qwen3_tts_tokenizer', 'qwen_tts.core', 'qwen_tts.cli', 'requests', 'pkg_resources.extern', 'backend.backends.mlx_backend', 'mlx', 'mlx.core', 'mlx.nn', 'mlx_audio', 'mlx_audio.tts', 'mlx_audio.stt']
 datas += collect_data_files('qwen_tts')
 datas += copy_metadata('qwen-tts')
 datas += copy_metadata('requests')
@@ -23,6 +23,16 @@ tmp_ret = collect_all('zipvoice')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 tmp_ret = collect_all('linacodec')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('lazy_loader')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('librosa')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('inflect')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('perth')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('piper_phonemize')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 tmp_ret = collect_all('mlx')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 tmp_ret = collect_all('mlx_audio')

docs/README.md

@@ -6,11 +6,7 @@ This is a Next.js application generated with
 Run development server:
 
 ```bash
-npm run dev
+bun run dev
-# or
-pnpm dev
-# or
-yarn dev
 ```
 
 Open http://localhost:3000 with your browser to see the result.

docs/app/(home)/layout.tsx

@@ -1,6 +0,0 @@
-import { HomeLayout } from 'fumadocs-ui/layouts/home';
-import { baseOptions } from '@/lib/layout.shared';
-
-export default function Layout({ children }: LayoutProps<'/'>) {
-  return <HomeLayout {...baseOptions()}>{children}</HomeLayout>;
-}

docs/app/(home)/page.tsx

@@ -1,5 +0,0 @@
-import { redirect } from 'next/navigation';
-
-export default function HomePage() {
-  redirect('/docs');
-}

docs/app/[[...slug]]/layout.tsx

@@ -1,8 +1,8 @@
-import { source } from '@/lib/source';
 import { DocsLayout } from 'fumadocs-ui/layouts/docs';
 import { baseOptions } from '@/lib/layout.shared';
+import { source } from '@/lib/source';
 
-export default function Layout({ children }: LayoutProps<'/docs'>) {
+export default function Layout({ children }: LayoutProps<'/[[...slug]]'>) {
   return (
     <DocsLayout tree={source.pageTree} {...baseOptions()}>
       {children}

docs/app/[[...slug]]/page.tsx

@@ -7,7 +7,7 @@ import { APIPage } from '@/components/api-page';
 import { getPageImage, source } from '@/lib/source';
 import { getMDXComponents } from '@/mdx-components';
 
-export default async function Page(props: PageProps<'/docs/[[...slug]]'>) {
+export default async function Page(props: PageProps<'/[[...slug]]'>) {
   const params = await props.params;
   const page = source.getPage(params.slug);
   if (!page) notFound();
@@ -59,7 +59,7 @@ export async function generateStaticParams() {
   return source.generateParams();
 }
 
-export async function generateMetadata(props: PageProps<'/docs/[[...slug]]'>): Promise<Metadata> {
+export async function generateMetadata(props: PageProps<'/[[...slug]]'>): Promise<Metadata> {
   const params = await props.params;
   const page = source.getPage(params.slug);
   if (!page) notFound();

docs/content/docs/api-reference/general/meta.json

@@ -0,0 +1,4 @@
+{
+  "title": "General",
+  "pages": ["root__get", "health_health_get"]
+}

docs/content/docs/api-reference/generation/meta.json

@@ -0,0 +1,8 @@
+{
+  "title": "Generation",
+  "pages": [
+    "generate_speech_generate_post",
+    "transcribe_audio_transcribe_post",
+    "get_audio_audio__generation_id__get"
+  ]
+}

docs/content/docs/api-reference/history/meta.json

@@ -0,0 +1,9 @@
+{
+  "title": "History",
+  "pages": [
+    "list_history_history_get",
+    "get_generation_history__generation_id__get",
+    "delete_generation_history__generation_id__delete",
+    "get_stats_history_stats_get"
+  ]
+}

docs/content/docs/api-reference/meta.json

@@ -1,5 +1,5 @@
 {
   "title": "API Reference",
   "defaultOpen": true,
-  "pages": ["unknown"]
+  "pages": ["general", "profiles", "generation", "history", "models"]
 }

docs/content/docs/api-reference/models/meta.json

@@ -0,0 +1,10 @@
+{
+  "title": "Models",
+  "pages": [
+    "get_model_status_models_status_get",
+    "load_model_models_load_post",
+    "unload_model_models_unload_post",
+    "trigger_model_download_models_download_post",
+    "get_model_progress_models_progress__model_name__get"
+  ]
+}

docs/content/docs/api-reference/profiles/meta.json

@@ -0,0 +1,13 @@
+{
+  "title": "Profiles",
+  "pages": [
+    "list_profiles_profiles_get",
+    "create_profile_profiles_post",
+    "get_profile_profiles__profile_id__get",
+    "update_profile_profiles__profile_id__put",
+    "delete_profile_profiles__profile_id__delete",
+    "get_profile_samples_profiles__profile_id__samples_get",
+    "add_profile_sample_profiles__profile_id__samples_post",
+    "delete_profile_sample_profiles_samples__sample_id__delete"
+  ]
+}

docs/content/docs/api-reference/unknown/meta.json

@@ -1,27 +0,0 @@
-{
-  "title": "Endpoints",
-  "pages": [
-    "root__get",
-    "health_health_get",
-    "list_profiles_profiles_get",
-    "create_profile_profiles_post",
-    "get_profile_profiles__profile_id__get",
-    "update_profile_profiles__profile_id__put",
-    "delete_profile_profiles__profile_id__delete",
-    "add_profile_sample_profiles__profile_id__samples_post",
-    "get_profile_samples_profiles__profile_id__samples_get",
-    "delete_profile_sample_profiles_samples__sample_id__delete",
-    "generate_speech_generate_post",
-    "list_history_history_get",
-    "get_generation_history__generation_id__get",
-    "delete_generation_history__generation_id__delete",
-    "get_stats_history_stats_get",
-    "transcribe_audio_transcribe_post",
-    "get_audio_audio__generation_id__get",
-    "load_model_models_load_post",
-    "unload_model_models_unload_post",
-    "get_model_progress_models_progress__model_name__get",
-    "get_model_status_models_status_get",
-    "trigger_model_download_models_download_post"
-  ]
-}

docs/content/docs/developer/building.mdx

@@ -159,12 +159,14 @@ Tauri looks for `voicebox-server-${PLATFORM}` in `src-tauri/binaries/` and bundl
 
 The `build-cuda-windows` job runs separately:
 
-1. Install PyTorch with CUDA 12.1
+1. Install PyTorch with CUDA 12.8
-2. Build with `build_binary.py --cuda`
+2. Build with `build_binary.py --cuda` (produces `--onedir` output)
-3. Split binary with `scripts/split_binary.py`
+3. Package with `scripts/package_cuda.py` into two archives:
-4. Upload parts as release artifacts
+   - `voicebox-server-cuda.tar.gz` — server core (~945 MB)
+   - `cuda-libs-cu128-v1.tar.gz` — NVIDIA runtime libraries (~1.7 GB, cached independently)
+4. Upload archives as release artifacts
 
-This binary is downloaded on-demand by users who enable CUDA in settings.
+This binary is downloaded on-demand by users who enable CUDA in settings. The CUDA libs archive is only re-downloaded when the CUDA toolkit version changes, not on every app update.
 
 ## Troubleshooting
 

docs/content/docs/developer/tts-engines.mdx

@@ -3,8 +3,12 @@ title: "TTS Engines"
 description: "How to add new text-to-speech engines to Voicebox"
 ---
 
+> **For humans:** This doc is optimized for AI agents to implement new TTS engines autonomously. It's structured as a phased workflow with explicit gates and a checklist so an agent can do the full integration — dependency research, backend, frontend, bundling — and hand you a draft release or prod build to test locally. It's also a useful reference if you're doing it yourself.
+
 Adding an engine touches ~10 files across 4 layers. The backend protocol work is straightforward — the real time sink is dependency hell, upstream library bugs, and PyInstaller bundling.
 
+**Do not start writing code until you complete Phase 0.** The v0.2.3 release was three patch releases of PyInstaller fixes because dependency research was skipped. Every issue — `inspect.getsource()` failures, missing native data files, metadata lookups, dtype mismatches — was discoverable by reading the model library's source code before integration began.
+
 ## Architecture Overview
 
 The backend is split into layers:
@@ -18,6 +22,133 @@ The backend is split into layers:
 
 New engines only need to touch `backends/` and `models.py` on the backend side — the route and service layers use a model config registry that handles dispatch automatically.
 
+## Phase 0: Dependency Research
+
+**This phase is mandatory.** Clone the model library and its key dependencies into a temporary directory and inspect them before writing any integration code. The goal is to produce a dependency audit that identifies every PyInstaller-incompatible pattern, every native data file, and every upstream bug you'll need to work around.
+
+### 0.1 Clone and Inspect the Model Library
+
+```bash
+# Create a throwaway workspace
+mkdir /tmp/engine-research && cd /tmp/engine-research
+
+# Clone the model library
+git clone https://github.com/org/model-library.git
+cd model-library
+```
+
+**Read these files first, in order:**
+
+1. **`setup.py` / `setup.cfg` / `pyproject.toml`** — Check pinned dependency versions. If the library pins `torch==2.6.0` or `numpy<1.26`, you'll need `--no-deps` installation and manual sub-dependency listing (this is what happened with `chatterbox-tts`).
+
+2. **`__init__.py` and the main model class** — Trace the import chain. Look for:
+   - `from_pretrained()` — does it call `huggingface_hub` internally? Does it pass `token=True` (which crashes without a stored HF token)?
+   - `from_local()` — does it exist? You may need manual `snapshot_download()` + `from_local()` to bypass download bugs.
+   - Device handling — does it default to CUDA? Does it support MPS? Many libraries crash on MPS with unsupported operators.
+
+3. **All `import` statements** — Recursively trace what the library imports. You're looking for:
+   - `inspect.getsource()` anywhere in the chain (search all `.py` files)
+   - `typeguard` / `@typechecked` decorators (these call `inspect.getsource()` at import time)
+   - `importlib.metadata.version()` or `pkg_resources.get_distribution()` (need `--copy-metadata`)
+   - `lazy_loader` (needs `--collect-all` to bundle `.pyi` stubs)
+
+### 0.2 Scan for PyInstaller-Incompatible Patterns
+
+Run these searches against the cloned library **and** its transitive dependencies:
+
+```bash
+# inspect.getsource — will crash in frozen binary without --collect-all
+grep -r "inspect.getsource\|getsource(" .
+
+# typeguard / @typechecked — calls inspect.getsource at import time
+grep -r "@typechecked\|from typeguard" .
+
+# importlib.metadata — needs --copy-metadata
+grep -r "importlib.metadata\|pkg_resources.get_distribution\|pkg_resources.require" .
+
+# Data files loaded at runtime — need --collect-all or --collect-data
+grep -r "Path(__file__).parent\|os.path.dirname(__file__)\|resources_path\|pkg_resources.resource_filename" .
+
+# Native library paths — may need env var override in frozen builds
+grep -r "/usr/share\|/usr/lib\|/usr/local\|espeak\|phonemize" .
+
+# torch.load without map_location — will crash on CPU-only builds
+grep -r "torch.load(" . | grep -v "map_location"
+
+# HuggingFace token bugs
+grep -r 'token=True\|token=os.getenv' .
+
+# Float64/Float32 assumptions — librosa returns float64, many models assume float32
+grep -r "torch.from_numpy\|\.double()\|float64" .
+
+# @torch.jit.script — calls inspect.getsource(), crashes in frozen builds
+grep -r "@torch.jit.script\|torch.jit.script" .
+
+# torchaudio.load — requires torchcodec in torchaudio 2.10+, use soundfile.read() instead
+grep -r "torchaudio.load\|torchaudio.save" .
+
+# Gated HuggingFace repos — models that hardcode gated repos as tokenizer/config sources
+grep -r "from_pretrained\|tokenizer_name\|AutoTokenizer" . | grep -i "llama\|meta-llama\|gated"
+```
+
+### 0.3 Install and Trace in a Throwaway Venv
+
+```bash
+# Create isolated venv
+python -m venv /tmp/engine-venv
+source /tmp/engine-venv/bin/activate
+
+# Install the package (try normally first)
+pip install model-package
+
+# Check if it conflicts with our stack
+pip install model-package torch==2.10 transformers==4.57.3 numpy>=1.26
+# If this fails, you need --no-deps:
+pip install --no-deps model-package
+
+# Get the full dependency tree
+pip show model-package  # Check Requires: field
+pip show -f model-package  # List all installed files (look for data files)
+
+# Check for non-PyPI dependencies
+pip install model-package 2>&1 | grep -i "no matching distribution"
+```
+
+### 0.4 Test Model Loading on CPU
+
+Before writing any integration code, verify the model works on CPU in a plain Python script:
+
+```python
+import torch
+# Force CPU to catch map_location bugs early
+model = ModelClass.from_pretrained("org/model", device="cpu")
+
+# Test with a float32 audio array (not float64)
+import numpy as np
+audio = np.random.randn(16000).astype(np.float32)
+output = model.generate("Hello world", audio)
+print(f"Output shape: {output.shape}, dtype: {output.dtype}, sample rate: {model.sample_rate}")
+```
+
+If this crashes, you've found a bug you'll need to monkey-patch. Common ones:
+- `RuntimeError: expected scalar type Float but found Double` → needs float32 cast
+- `RuntimeError: map_location` → needs `torch.load` patch
+- `RuntimeError: Unsupported operator aten::...` → needs MPS skip
+
+### 0.5 Produce a Dependency Audit
+
+Before proceeding to Phase 1, write down:
+
+1. **PyPI vs non-PyPI deps** — which packages need `--find-links`, `git+https://`, or `--no-deps`?
+2. **PyInstaller directives needed** — which packages need `--collect-all`, `--copy-metadata`, `--hidden-import`?
+3. **Runtime data files** — which packages ship data files (YAML, pretrained weights, phoneme tables, shader libraries) that must be bundled?
+4. **Native library paths** — which packages look for data at system paths that won't exist in a frozen binary?
+5. **Monkey-patches needed** — `torch.load` map_location, float64→float32 casts, MPS skip, HF token bypass, etc.
+6. **Sample rate** — what does the engine output? (24kHz, 44.1kHz, 48kHz)
+7. **Model download method** — `from_pretrained()` with library-managed download, or manual `snapshot_download()` + `from_local()`?
+
+This audit becomes your implementation plan for Phases 1, 4, and 5.
+
 ## Phase 1: Backend Implementation
 
 ### 1.1 Create the Backend File
@@ -148,6 +279,8 @@ In `app/src/lib/hooks/useGenerationForm.ts`:
 - Add engine-to-model-name mapping
 - Update payload construction for engine-specific fields
 
+**Watch out for model naming inconsistencies.** The HuggingFace repo name, the model size label, and the API model name don't always follow predictable patterns. For example, TADA's 3B model is named `tada-3b-ml` (not `tada-3b`), because it's a multilingual variant. Always check the actual repo names and build the frontend model name mapping from those, not from assumptions like `{engine}-{size}`.
+
 ### 3.5 Model Management
 
 In `app/src/components/ServerSettings/ModelManagement.tsx`:
@@ -155,54 +288,173 @@ In `app/src/components/ServerSettings/ModelManagement.tsx`:
 
 ## Phase 4: Dependencies
 
+Use the dependency audit from Phase 0 to drive this phase. You should already know what packages are needed, which conflict, and which require special installation.
+
 ### 4.1 Python Dependencies
 
-Add to `backend/requirements.txt`. Watch for:
+Add to `backend/requirements.txt`. There are three installation patterns, depending on what Phase 0 revealed:
 
-**Pinned dependency conflicts** — If the model package pins old versions, install with `--no-deps`:
+**Normal PyPI packages:**
+```
+some-model-package>=1.0.0
+```
+
+**Pinned dependency conflicts (`--no-deps`)** — If the model package pins old versions of torch/numpy/transformers, install with `--no-deps` and list sub-dependencies manually. This is the pattern used for `chatterbox-tts`:
 ```bash
+# In justfile / CI setup:
 pip install --no-deps chatterbox-tts
+
+# In requirements.txt — list each actual sub-dependency:
+conformer>=0.3.2
+diffusers>=0.31.0
+omegaconf>=2.3.0
+resemble-perth>=0.0.2
+s3tokenizer>=0.1.6
 ```
 
-Then list sub-dependencies manually in `requirements.txt`.
+To identify sub-deps: `pip show chatterbox-tts` → `Requires:` field, then cross-reference against existing `requirements.txt` to avoid duplicates.
 
-**Non-PyPI packages:**
+**Non-PyPI packages** — Some libraries only exist on GitHub or require custom indexes:
-```
-linacodec @ git+https://github.com/user/repo.git
 ```
+# Git-only packages (no PyPI release)
+linacodec @ git+https://github.com/ysharma3501/LinaCodec.git
+Zipvoice @ git+https://github.com/ysharma3501/LuxTTS.git
 
-**Custom package indexes:**
+# Custom package indexes (C extensions with platform-specific wheels)
-```
 --find-links https://k2-fsa.github.io/icefall/piper_phonemize.html
+piper-phonemize>=1.2.0
 ```
 
-### 4.2 Identifying Hidden Sub-Dependencies
+### 4.2 Dependency Conflict Resolution
 
-1. Install the package normally in a throwaway venv
+Check for conflicts with the existing stack before adding anything:
-2. Run `pip show <package>` to get its `Requires:` list
-3. Cross-reference against existing requirements.txt
-4. Test that the engine loads and generates
 
-## Phase 5: PyInstaller Bundling
+```bash
+# Our current stack pins (approximate):
+# Python 3.12+, torch>=2.10, transformers>=4.57, numpy>=1.26
 
-This is where most of the pain lives. Common issues:
+# Test compatibility
+pip install model-package torch==2.10 transformers==4.57.3 numpy>=1.26
 
-| Issue | Symptom | Fix |
+# If it fails, check what the package pins:
-|-------|---------|-----|
+pip show model-package | grep Requires
-| `inspect.getsource()` at import | "could not get source code" | `--collect-all <package>` |
+# Look at setup.py/pyproject.toml for version constraints
-| Data files (yaml, .pth.tar) | FileNotFoundError at runtime | `--collect-all <package>` |
+```
-| Native data paths (espeak-ng) | Library looks at `/usr/share/...` | Set env var in frozen builds |
-| `importlib.metadata` lookups | "No package metadata found" | `--copy-metadata <package>` |
-| Dynamic imports | ModuleNotFoundError | `--hidden-import <module>` |
 
-### Testing Frozen Builds
+**Known incompatible patterns in the wild:**
+- `torch==2.6.0` — many older packages pin this
+- `numpy<1.26` — conflicts with Python 3.12+
+- `transformers==4.46.3` — many packages pin old transformers
+- `onnxruntime` pinned versions — often conflict with torch
 
-You can't skip this. Models that work in `python -m uvicorn` will break in the PyInstaller binary.
+### 4.3 Update Installation Scripts
+
+Dependencies must be added in multiple places:
+
+| File | What to add |
+|------|------------|
+| `backend/requirements.txt` | Package and version constraint |
+| `justfile` | `--no-deps` install line if needed (in `setup-python` and `setup-python-release` targets) |
+| `.github/workflows/release.yml` | Same `--no-deps` line in CI build steps |
+| `Dockerfile` | Same install commands for Docker builds |
+
+## Phase 5: PyInstaller Bundling (`build_binary.py`)
+
+This is where most of the pain lives. **The v0.2.3 release was entirely dedicated to fixing bundling issues** — every new engine that shipped in v0.2.1 (LuxTTS, Chatterbox, Chatterbox Turbo) worked in dev but failed in production builds. Don't skip this phase.
+
+### 5.1 Register Your Engine in `build_binary.py`
+
+Every new engine needs entries in `backend/build_binary.py`. This file drives PyInstaller and is the single most common source of "works in dev, breaks in prod" bugs. You need to decide which PyInstaller directives your engine's dependencies require:
+
+| Directive | What It Does | When You Need It |
+|-----------|-------------|-----------------|
+| `--hidden-import <module>` | Includes a module PyInstaller can't detect via static analysis | Dynamic imports, lazy imports, plugin architectures |
+| `--collect-all <package>` | Bundles source `.py` files, data files, AND native libraries | Packages that call `inspect.getsource()` at import time (e.g. `inflect` via `typeguard`'s `@typechecked`), or that ship pretrained model files (e.g. `perth` ships `.pth.tar` + `hparams.yaml`) |
+| `--collect-data <package>` | Bundles only data files (not source or native libs) | Packages with YAML configs, vocab files, etc. |
+| `--collect-submodules <package>` | Bundles all submodules | Packages with deep module trees that PyInstaller misses |
+| `--copy-metadata <package>` | Copies `importlib.metadata` info | Packages that call `importlib.metadata.version()` or `pkg_resources.get_distribution()` at runtime. Already required for: `requests`, `transformers`, `huggingface-hub`, `tokenizers`, `safetensors`, `tqdm` |
+
+**Example: adding hidden imports and collect-all for a new engine:**
+
+```python
+# In build_binary.py, inside the args list:
+"--hidden-import",
+"backend.backends.your_engine_backend",
+"--hidden-import",
+"your_engine_package",
+"--hidden-import",
+"your_engine_package.inference",
+"--collect-all",
+"some_dependency_that_uses_inspect_getsource",
+"--copy-metadata",
+"some_dependency_that_checks_its_own_version",
+```
+
+### 5.2 Lessons from v0.2.3 — Real Failures and Their Fixes
+
+These are actual production failures from shipping new engines. Every one of these passed `python -m uvicorn` in dev:
+
+| Engine | Failure | Root Cause | Fix |
+|--------|---------|-----------|-----|
+| LuxTTS | `"could not get source code"` on import | `inflect` uses `typeguard`'s `@typechecked` which calls `inspect.getsource()` — needs `.py` source files, not just bytecode | `--collect-all inflect` |
+| LuxTTS | `espeak-ng-data` not found | `piper_phonemize` C library looks for data at `/usr/share/espeak-ng-data/` which doesn't exist in the bundle | `--collect-all piper_phonemize` + set `ESPEAK_DATA_PATH` env var at runtime (see 5.3) |
+| LuxTTS | `inspect.getsource` error in Vocos codec | `linacodec` and `zipvoice` use source introspection | `--collect-all linacodec` + `--collect-all zipvoice` |
+| Chatterbox | `FileNotFoundError` for watermark model | `perth` ships pretrained model files (`hparams.yaml`, `.pth.tar`) that PyInstaller doesn't bundle by default | `--collect-all perth` |
+| All engines | `importlib.metadata` failures | Frozen binary doesn't include package metadata for `huggingface-hub`, `transformers`, etc. | `--copy-metadata` for each affected package |
+| All engines | Download progress bars stuck at 0% | `huggingface_hub` silently disables tqdm progress bars based on logger level in frozen builds — our progress tracker never receives byte updates | Force-enable tqdm's internal counter in `HFProgressTracker` |
+| TADA | `inspect.getsource` error in DAC's `Snake1d` | `@torch.jit.script` calls `inspect.getsource()` which fails without `.py` source files | Wrote a lightweight shim (`dac_shim.py`) reimplementing `Snake1d` without `@torch.jit.script`, registered fake `dac.*` modules in `sys.modules` |
+| All engines | `NameError: name 'obj' is not defined` on macOS | Python 3.12.0 has a [CPython bug](https://github.com/pyinstaller/pyinstaller/issues/7992) that corrupts bytecode when PyInstaller rewrites code objects | Upgrade to Python 3.12.13+ |
+| All engines | `resource_tracker` subprocess crash | `multiprocessing` in frozen binaries needs `freeze_support()` called before anything else | Added to `server.py` entry point |
+
+### 5.3 Runtime Frozen-Build Handling (`server.py`)
+
+Some fixes can't live in `build_binary.py` — they need runtime detection. The entry point `backend/server.py` handles these before any heavy imports:
+
+```python
+# 1. freeze_support() — MUST be called before any multiprocessing use
+import multiprocessing
+multiprocessing.freeze_support()
+
+# 2. Native data paths — redirect C libraries to bundled data
+if getattr(sys, 'frozen', False):
+    _meipass = getattr(sys, '_MEIPASS', os.path.dirname(sys.executable))
+    _espeak_data = os.path.join(_meipass, 'piper_phonemize', 'espeak-ng-data')
+    if os.path.isdir(_espeak_data):
+        os.environ.setdefault('ESPEAK_DATA_PATH', _espeak_data)
+
+# 3. stdout/stderr safety — PyInstaller --noconsole on Windows sets these to None
+if not _is_writable(sys.stdout):
+    sys.stdout = open(os.devnull, 'w')
+```
+
+If your engine's dependencies include native libraries that look for data at system paths (like espeak-ng does), you'll need to add a similar `os.environ.setdefault()` block here.
+
+### 5.4 CUDA vs CPU Build Branching
+
+`build_binary.py` produces two different binaries:
+
+- **`voicebox-server`** (CPU) — excludes all `nvidia.*` packages to avoid bundling ~3 GB of CUDA DLLs
+- **`voicebox-server-cuda`** — includes `torch.cuda` and `torch.backends.cudnn`
+
+On Windows, if the build environment has CUDA torch installed but you're building the CPU binary, the script temporarily swaps to CPU-only torch and restores CUDA torch afterward. This prevents PyInstaller from accidentally bundling CUDA libraries into the CPU build.
+
+New engine imports go in the **common section** (not the CUDA or MLX conditional blocks) unless your engine has platform-specific dependencies.
+
+### 5.5 MLX Conditional Inclusion
+
+Apple Silicon builds conditionally include MLX hidden imports and `--collect-all mlx` / `--collect-all mlx_audio`. If your engine has an MLX-specific backend variant, add its imports inside the `if is_apple_silicon() and not cuda:` block.
+
+### 5.6 Testing Frozen Builds
+
+You can't skip this. Models that work in `python -m uvicorn` will break in the PyInstaller binary. The v0.2.3 release required **three patch releases** (v0.2.1 → v0.2.2 → v0.2.3) to get all engines working in production.
 
 1. Build: `just build`
-2. Run and try download + load + generate
+2. Launch the binary directly (not via `python -m`)
-3. Check stderr for the actual error
+3. Test the **full chain**: download → load → generate → progress tracking
-4. Fix, rebuild, repeat
+4. Check stderr for the actual error (logs go to stderr for Tauri sidecar capture)
+5. Fix, rebuild, repeat
+
+**Common gotcha:** testing only generation with a pre-cached model from your dev install. Always test with a clean model cache to verify downloads work too.
 
 ## Phase 6: Common Upstream Workarounds
 
@@ -240,6 +492,90 @@ def _get_device(self):
     return "cpu"  # Skip MPS
 ```
 
+### Gated HuggingFace repos as hardcoded config sources
+
+Some models hardcode a gated HuggingFace repo as their tokenizer or config source (e.g., TADA hardcodes `"meta-llama/Llama-3.2-1B"` in both its `AlignerConfig` and `TadaConfig`). This silently fails without HF authentication.
+
+**Fix:** Download from an ungated mirror and patch the config objects directly:
+
+```python
+# Download tokenizer from ungated mirror
+UNGATED_TOKENIZER = "unsloth/Llama-3.2-1B"
+tokenizer_path = snapshot_download(UNGATED_TOKENIZER, token=None)
+
+# Patch the model config to use the local path instead of the gated repo
+config = ModelConfig.from_pretrained(model_path)
+config.tokenizer_name = tokenizer_path
+model = ModelClass.from_pretrained(model_path, config=config)
+```
+
+**Do NOT monkey-patch `AutoTokenizer.from_pretrained`** — it's a classmethod, and replacing it corrupts the descriptor, which breaks other engines that use different tokenizers (e.g., Qwen uses a Qwen tokenizer via `AutoTokenizer`). Always patch at the config level, not the class method level.
+
+### `torchaudio.load()` requires `torchcodec` in 2.10+
+
+As of `torchaudio>=2.10`, `torchaudio.load()` requires the `torchcodec` package for audio I/O. If your engine or backend code uses `torchaudio.load()`, replace it with `soundfile`:
+
+```python
+# Before (breaks without torchcodec):
+import torchaudio
+waveform, sr = torchaudio.load("audio.wav")
+
+# After:
+import soundfile as sf
+import torch
+data, sr = sf.read("audio.wav", dtype="float32")
+waveform = torch.from_numpy(data).unsqueeze(0)
+```
+
+Note: `torchaudio.functional.resample()` and other pure-PyTorch math functions work fine without `torchcodec` — only the I/O functions are affected.
+
+### `@torch.jit.script` breaks in frozen builds
+
+`torch.jit.script` calls `inspect.getsource()` to parse the decorated function's source code. In a PyInstaller binary, `.py` source files aren't available, so this crashes at import time.
+
+**Fix:** Remove or avoid `@torch.jit.script` decorators. If the decorated function comes from an upstream dependency, write a shim that reimplements the function without the decorator (see "Toxic dependency chains" below).
+
+### Toxic dependency chains — the shim pattern
+
+Sometimes a model library depends on a package with a massive, hostile transitive dependency tree, but only uses a tiny piece of it. When the dependency chain is unbuildable or would pull in dozens of unwanted packages, the right move is to write a lightweight shim.
+
+**Example:** TADA depends on `descript-audio-codec` (DAC), which pulls in `descript-audiotools` -> `onnx`, `tensorboard`, `protobuf`, `matplotlib`, `pystoi`, etc. The `onnx` package fails to build from source on macOS. But TADA only uses `Snake1d` from DAC — a 7-line PyTorch module.
+
+**Solution:** Create a shim at `backend/utils/dac_shim.py` that registers fake modules in `sys.modules`:
+
+```python
+import sys
+import types
+import torch
+from torch import nn
+
+def snake(x, alpha):
+    """Snake activation — reimplemented without @torch.jit.script."""
+    return x + (1.0 / (alpha + 1e-9)) * torch.sin(alpha * x).pow(2)
+
+class Snake1d(nn.Module):
+    def __init__(self, channels):
+        super().__init__()
+        self.alpha = nn.Parameter(torch.ones(1, channels, 1))
+    def forward(self, x):
+        return snake(x, self.alpha)
+
+# Register fake dac.* modules so "from dac.nn.layers import Snake1d" works
+_nn = types.ModuleType("dac.nn")
+_layers = types.ModuleType("dac.nn.layers")
+_layers.Snake1d = Snake1d
+_nn.layers = _layers
+
+for name, mod in [("dac", types.ModuleType("dac")),
+                   ("dac.nn", _nn), ("dac.nn.layers", _layers)]:
+    sys.modules[name] = mod
+```
+
+**Key rules for shims:**
+- Import the shim **before** importing the model library (so it finds the fake modules first)
+- Do NOT use `@torch.jit.script` in the shim (see above)
+- Only reimplement what the model actually uses — check the import chain carefully
+
 ## Upcoming Engines
 
 Based on the current model landscape, these are candidates for future integration:
@@ -250,8 +586,83 @@ Based on the current model landscape, these are candidates for future integratio
 | **Fish Speech** | 50+ | Medium | Word-level control via inline text | Ready |
 | **Kokoro-82M** | English | 82M | CPU realtime, Apache 2.0 | Ready |
 | **XTTS-v2** | 17+ | Medium | Zero-shot cloning | Ready |
-| **HumeAI TADA** | EN (1B), Multi (3B) | Medium | 700s+ coherent audio, synced transcripts | Needs vetting |
 | **MOSS-TTS** | Multilingual | Medium | Text-to-voice design, multi-speaker dialogue | Needs vetting |
 | **Pocket TTS** | English | ~100M | CPU-first, >1× realtime | Needs vetting |
 
-The multi-engine architecture is now in place, making new model integration straightforward (~1 day for a well-documented model with a PyPI package).
+## Implementation Checklist
+
+Use this as a gate between phases. Do not proceed to the next phase until every item in the current phase is checked.
+
+### Phase 0: Dependency Research
+- [ ] Cloned model library source into a temp directory
+- [ ] Read `setup.py` / `pyproject.toml` — noted pinned dependency versions
+- [ ] Traced all imports from the model class through to leaf dependencies
+- [ ] Searched for `inspect.getsource`, `@typechecked`, `typeguard` in the full dependency tree
+- [ ] Searched for `importlib.metadata`, `pkg_resources.get_distribution` in the dependency tree
+- [ ] Searched for `Path(__file__).parent`, `os.path.dirname(__file__)`, hardcoded system paths
+- [ ] Searched for `torch.load` calls missing `map_location`
+- [ ] Searched for `torch.from_numpy` without `.float()` cast
+- [ ] Searched for `token=True` or `token=os.getenv("HF_TOKEN")` in HuggingFace calls
+- [ ] Searched for `@torch.jit.script` / `torch.jit.script` (crashes in frozen builds)
+- [ ] Searched for `torchaudio.load` / `torchaudio.save` (requires `torchcodec` in 2.10+)
+- [ ] Searched for hardcoded gated HuggingFace repo names (e.g., `meta-llama/*`)
+- [ ] Evaluated whether any dependency is used minimally enough to shim instead of install
+- [ ] Tested model loading and generation on CPU in a throwaway venv
+- [ ] Tested with a clean HuggingFace cache (no pre-downloaded models)
+- [ ] Produced a written dependency audit documenting all findings
+
+### Phase 1: Backend Implementation
+- [ ] Created `backend/backends/<engine>_backend.py` implementing `TTSBackend` protocol
+- [ ] Chose voice prompt pattern (pre-computed tensors vs deferred file paths)
+- [ ] Implemented all monkey-patches identified in Phase 0
+- [ ] Used `get_torch_device()` from `backends/base.py` for device selection
+- [ ] Used `model_load_progress()` from `backends/base.py` for download/load tracking
+- [ ] Tested: model downloads correctly
+- [ ] Tested: model loads on CPU
+- [ ] Tested: generation produces valid audio
+- [ ] Tested: voice cloning from reference audio works
+- [ ] Registered `ModelConfig` in `backends/__init__.py`
+- [ ] Added to `TTS_ENGINES` dict
+- [ ] Added factory branch in `get_tts_backend_for_engine()`
+- [ ] Updated engine regex in `backend/models.py`
+
+### Phase 2–3: Route, Service, and Frontend
+- [ ] Confirmed zero changes needed in routes/services (or documented why custom behavior is needed)
+- [ ] Added engine to TypeScript union type in `app/src/lib/api/types.ts`
+- [ ] Added language map entry in `app/src/lib/constants/languages.ts`
+- [ ] Added to `ENGINE_OPTIONS` and `ENGINE_DESCRIPTIONS` in `EngineModelSelector.tsx`
+- [ ] Added to Zod schema and model-name mapping in `useGenerationForm.ts`
+- [ ] Added description in `ModelManagement.tsx`
+
+### Phase 4: Dependencies
+- [ ] Added packages to `backend/requirements.txt`
+- [ ] If `--no-deps` needed: listed sub-dependencies explicitly
+- [ ] If git-only packages: added `@ git+https://...` entries
+- [ ] If custom index needed: added `--find-links` line
+- [ ] Updated `justfile` setup targets
+- [ ] Updated `.github/workflows/release.yml` build steps
+- [ ] Updated `Dockerfile` if applicable
+- [ ] Verified `pip install` succeeds in a clean venv with existing requirements
+
+### Phase 5: PyInstaller Bundling
+- [ ] Added `--hidden-import` entries in `build_binary.py` for:
+  - [ ] `backend.backends.<engine>_backend`
+  - [ ] The model package and its key submodules
+- [ ] Added `--collect-all` for any packages that:
+  - [ ] Use `inspect.getsource()` / `@typechecked`
+  - [ ] Ship pretrained model data files (`.pth.tar`, `.yaml`, etc.)
+  - [ ] Ship native data files (phoneme tables, shader libraries, etc.)
+- [ ] Added `--copy-metadata` for any packages that use `importlib.metadata`
+- [ ] If engine has native data paths: added `os.environ.setdefault()` in `server.py`
+- [ ] Built frozen binary with `just build`
+- [ ] Tested in frozen binary with **clean model cache** (not pre-cached from dev):
+  - [ ] Model download works with real-time progress
+  - [ ] Model loading works
+  - [ ] Generation produces valid audio
+  - [ ] No errors in stderr logs
+
+### Phase 6: Final Verification
+- [ ] Engine works in dev mode (`just dev`)
+- [ ] Engine works in frozen binary (`just build` → run binary directly)
+- [ ] Tested on target platform (macOS for MLX, Windows/Linux for CUDA)
+- [ ] No regressions in existing engines