Fix R2 bucket name: voicebox (not voicebox-releases)

- Update CI workflow to upload CUDA binary to R2 instead of GitHub
- Use custom domain: downloads.voicebox.sh
- Update release notes with R2 download link
- Add R2 setup documentation with GitHub secrets instructions
- Add local test script for R2 uploads
- Fixes GitHub 2GB release asset limit issue

Required GitHub secrets:
- R2_ACCESS_KEY_ID
- R2_SECRET_ACCESS_KEY
- R2_ENDPOINT

Cost: ~$0.04/month (free bandwidth with R2)

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.1.11
+current_version = 0.1.12
 commit = True
 tag = True
 tag_name = v{new_version}

.github/workflows/release.yml

@@ -72,12 +72,18 @@ jobs:
           chmod +x scripts/build-server.sh
           ./scripts/build-server.sh
 
-      - name: Build Python server (Windows)
+      - name: Build CPU Python server (Windows)
         if: matrix.platform == 'windows-latest'
         shell: bash
         run: |
           cd backend
-          python build_binary.py
+
+          echo "Installing CPU-only PyTorch..."
+          pip uninstall -y torch torchvision torchaudio
+          pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
+
+          echo "Building CPU server binary..."
+          python build_binary.py cpu
 
           # Get platform tuple
           PLATFORM=$(rustc --print host-tuple)
@@ -85,9 +91,31 @@ jobs:
           # Create binaries directory
           mkdir -p ../tauri/src-tauri/binaries
 
-          # Copy with platform suffix
+          # Copy CPU version (default for installer)
           cp dist/voicebox-server.exe ../tauri/src-tauri/binaries/voicebox-server-${PLATFORM}.exe
-          echo "Built voicebox-server-${PLATFORM}.exe"
+          echo "Built CPU server: voicebox-server-${PLATFORM}.exe (~500MB)"
+
+      - name: Build CUDA Python server (Windows)
+        if: matrix.platform == 'windows-latest'
+        shell: bash
+        run: |
+          cd backend
+
+          echo "Installing CUDA PyTorch..."
+          pip uninstall -y torch torchvision torchaudio
+          pip install torch --index-url https://download.pytorch.org/whl/cu121 --force-reinstall --no-deps
+          pip install torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
+
+          echo "Building CUDA server binary..."
+          python build_binary.py cuda
+
+          # Get platform tuple
+          PLATFORM=$(rustc --print host-tuple)
+
+          # Copy CUDA version for separate upload
+          mkdir -p cuda-release
+          cp dist/voicebox-server-cuda.exe cuda-release/voicebox-server-cuda-${PLATFORM}.exe
+          echo "Built CUDA server: voicebox-server-cuda-${PLATFORM}.exe (~3GB)"
 
       - name: Setup Bun
         uses: oven-sh/setup-bun@v2
@@ -144,11 +172,41 @@ jobs:
             ### 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
+            - **Windows**: Download the `.msi` installer - includes CPU-only inference (~500MB)
             - **Linux**: Download the `.AppImage` or `.deb` package
 
+            ### NVIDIA GPU Acceleration (Windows)
+            Windows users with NVIDIA GPUs can enable CUDA for 4-5x faster inference:
+            1. Install the app normally (CPU version included in installer)
+            2. The app will detect your GPU and offer to download CUDA support automatically
+            3. Or manually download: [voicebox-server-cuda-x86_64-pc-windows-msvc.exe](https://downloads.voicebox.sh/cuda/__VERSION__/voicebox-server-cuda-x86_64-pc-windows-msvc.exe) (~2.4GB)
+
             The app includes automatic updates - future updates will be installed automatically.
           releaseDraft: true
           prerelease: false
           args: ${{ matrix.args }}
           includeUpdaterJson: true
+
+      - name: Upload CUDA server to Cloudflare R2 (Windows only)
+        if: matrix.platform == 'windows-latest'
+        env:
+          AWS_ACCESS_KEY_ID: ${{ secrets.R2_ACCESS_KEY_ID }}
+          AWS_SECRET_ACCESS_KEY: ${{ secrets.R2_SECRET_ACCESS_KEY }}
+          R2_ENDPOINT: ${{ secrets.R2_ENDPOINT }}
+        run: |
+          # Install AWS CLI if not available
+          pip install awscli
+
+          # Get version from tag
+          VERSION=${GITHUB_REF#refs/tags/}
+
+          # Get platform tuple
+          PLATFORM=$(rustc --print host-tuple)
+
+          # Upload to R2
+          aws s3 cp backend/cuda-release/voicebox-server-cuda-${PLATFORM}.exe \
+            s3://voicebox/cuda/${VERSION}/voicebox-server-cuda-${PLATFORM}.exe \
+            --endpoint-url $R2_ENDPOINT \
+            --acl public-read
+
+          echo "CUDA binary uploaded to: https://downloads.voicebox.sh/cuda/${VERSION}/voicebox-server-cuda-${PLATFORM}.exe"

app/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@voicebox/app",
-  "version": "0.1.11",
+  "version": "0.1.12",
   "private": true,
   "type": "module",
   "scripts": {

app/src/App.tsx

@@ -1,13 +1,14 @@
-import { useEffect, useRef, useState } from 'react';
 import { RouterProvider } from '@tanstack/react-router';
+import { useEffect, useRef, useState } from 'react';
 import voiceboxLogo from '@/assets/voicebox-logo.png';
 import ShinyText from '@/components/ShinyText';
 import { TitleBarDragRegion } from '@/components/TitleBarDragRegion';
+import { useAutoUpdater } from '@/hooks/useAutoUpdater';
 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 { useServerStore } from '@/stores/serverStore';
-import { usePlatform } from '@/platform/PlatformContext';
 
 const LOADING_MESSAGES = [
   'Warming up tensors...',
@@ -38,6 +39,9 @@ function App() {
   const [loadingMessageIndex, setLoadingMessageIndex] = useState(0);
   const serverStartingRef = useRef(false);
 
+  // Automatically check for app updates on startup and show toast notifications
+  useAutoUpdater({ checkOnMount: true, showToast: true });
+
   // Sync stored setting to Rust on startup
   useEffect(() => {
     if (platform.metadata.isTauri) {
@@ -46,14 +50,18 @@ function App() {
         console.error('Failed to sync initial setting to Rust:', error);
       });
     }
-  }, [platform]);
+    // Empty dependency array - platform is stable from context, only run once
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.metadata.isTauri, platform.lifecycle]);
 
   // Setup lifecycle callbacks
   useEffect(() => {
     platform.lifecycle.onServerReady = () => {
       setServerReady(true);
     };
-  }, [platform]);
+    // Empty dependency array - platform is stable from context, only run once
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.lifecycle]);
 
   // Setup window close handler and auto-start server when running in Tauri (production only)
   useEffect(() => {
@@ -111,7 +119,9 @@ function App() {
       // Window close event handles server shutdown based on setting
       serverStartingRef.current = false;
     };
-  }, [platform]);
+    // Empty dependency array - platform is stable from context, only run once
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.metadata.isTauri, platform.lifecycle]);
 
   // Cycle through loading messages every 3 seconds
   useEffect(() => {

app/src/components/History/HistoryTable.tsx

@@ -1,6 +1,13 @@
-import { AudioWaveform, Download, FileArchive, Loader2, MoreHorizontal, Play, Trash2 } from 'lucide-react';
+import {
+  AudioWaveform,
+  Download,
+  FileArchive,
+  Loader2,
+  MoreHorizontal,
+  Play,
+  Trash2,
+} from 'lucide-react';
 import { useEffect, useRef, useState } from 'react';
-import type { HistoryResponse } from '@/lib/api/types';
 import { Button } from '@/components/ui/button';
 import {
   Dialog,
@@ -19,6 +26,7 @@ import {
 import { Textarea } from '@/components/ui/textarea';
 import { useToast } from '@/components/ui/use-toast';
 import { apiClient } from '@/lib/api/client';
+import type { HistoryResponse } from '@/lib/api/types';
 import { BOTTOM_SAFE_AREA_PADDING } from '@/lib/constants/ui';
 import {
   useDeleteGeneration,
@@ -50,7 +58,11 @@ export function HistoryTable() {
   const limit = 20;
   const { toast } = useToast();
 
-  const { data: historyData, isLoading, isFetching } = useHistory({
+  const {
+    data: historyData,
+    isLoading,
+    isFetching,
+  } = useHistory({
     limit,
     offset: page * limit,
   });
@@ -280,6 +292,7 @@ export function HistoryTable() {
                     <Textarea
                       value={gen.text}
                       className="flex-1 resize-none text-sm text-muted-foreground select-text"
+                      readOnly
                     />
                   </div>
 

app/src/components/ServerSettings/ModelManagement.tsx

@@ -1,6 +1,6 @@
 import { useMutation, useQuery, useQueryClient } from '@tanstack/react-query';
 import { Download, Loader2, Trash2 } from 'lucide-react';
-import { useState } from 'react';
+import { useCallback, useState } from 'react';
 import {
   AlertDialog,
   AlertDialogAction,
@@ -17,7 +17,6 @@ import { Card, CardContent, CardDescription, CardHeader, CardTitle } from '@/com
 import { useToast } from '@/components/ui/use-toast';
 import { apiClient } from '@/lib/api/client';
 import { useModelDownloadToast } from '@/lib/hooks/useModelDownloadToast';
-import { ModelProgress } from './ModelProgress';
 
 export function ModelManagement() {
   const { toast } = useToast();
@@ -27,15 +26,36 @@ export function ModelManagement() {
 
   const { data: modelStatus, isLoading } = useQuery({
     queryKey: ['modelStatus'],
-    queryFn: () => apiClient.getModelStatus(),
+    queryFn: async () => {
+      console.log('[Query] Fetching model status');
+      const result = await apiClient.getModelStatus();
+      console.log('[Query] Model status fetched:', result);
+      return result;
+    },
     refetchInterval: 5000, // Refresh every 5 seconds
   });
 
+  // Callbacks for download completion
+  const handleDownloadComplete = useCallback(() => {
+    console.log('[ModelManagement] Download complete, clearing state');
+    setDownloadingModel(null);
+    setDownloadingDisplayName(null);
+    queryClient.invalidateQueries({ queryKey: ['modelStatus'] });
+  }, [queryClient]);
+
+  const handleDownloadError = useCallback(() => {
+    console.log('[ModelManagement] Download error, clearing state');
+    setDownloadingModel(null);
+    setDownloadingDisplayName(null);
+  }, []);
+
   // Use progress toast hook for the downloading model
   useModelDownloadToast({
     modelName: downloadingModel || '',
     displayName: downloadingDisplayName || '',
     enabled: !!downloadingModel && !!downloadingDisplayName,
+    onComplete: handleDownloadComplete,
+    onError: handleDownloadError,
   });
 
   const [deleteDialogOpen, setDeleteDialogOpen] = useState(false);
@@ -45,44 +65,69 @@ export function ModelManagement() {
     sizeMb?: number;
   } | null>(null);
 
-  const downloadMutation = useMutation({
+  const handleDownload = async (modelName: string) => {
-    mutationFn: (modelName: string) => {
+    console.log('[Download] Button clicked for:', modelName, 'at', new Date().toISOString());
-      setDownloadingModel(modelName);
+    
-      // Find display name from model status
+    // Find display name
     const model = modelStatus?.models.find((m) => m.model_name === modelName);
-      setDownloadingDisplayName(model?.display_name || modelName);
+    const displayName = model?.display_name || modelName;
-      return apiClient.triggerModelDownload(modelName);
+    
-    },
+    try {
-    onSuccess: () => {
+      // IMPORTANT: Call the API FIRST before setting state
-      // Download completed - clear state and refetch status
+      // Setting state enables the SSE EventSource in useModelDownloadToast,
-      setDownloadingModel(null);
+      // which can block/delay the download fetch due to HTTP/1.1 connection limits
-      setDownloadingDisplayName(null);
+      console.log('[Download] Calling download API for:', modelName);
+      const result = await apiClient.triggerModelDownload(modelName);
+      console.log('[Download] Download API responded:', result);
+      
+      // NOW set state to enable SSE tracking (after download has started on backend)
+      setDownloadingModel(modelName);
+      setDownloadingDisplayName(displayName);
+      
+      // Download initiated successfully - state will be cleared when SSE reports completion
+      // or by the polling interval detecting the model is downloaded
       queryClient.invalidateQueries({ queryKey: ['modelStatus'] });
-    },
+    } catch (error) {
-    onError: (error: Error) => {
+      console.error('[Download] Download failed:', error);
       setDownloadingModel(null);
       setDownloadingDisplayName(null);
       toast({
         title: 'Download failed',
-        description: error.message,
+        description: error instanceof Error ? error.message : 'Unknown error',
         variant: 'destructive',
       });
-    },
+    }
-  });
+  };
 
   const deleteMutation = useMutation({
-    mutationFn: (modelName: string) => apiClient.deleteModel(modelName),
+    mutationFn: async (modelName: string) => {
-    onSuccess: () => {
+      console.log('[Delete] Deleting model:', modelName);
+      const result = await apiClient.deleteModel(modelName);
+      console.log('[Delete] Model deleted successfully:', modelName);
+      return result;
+    },
+    onSuccess: async (_data, _modelName) => {
+      console.log('[Delete] onSuccess - showing toast and invalidating queries');
       toast({
         title: 'Model deleted',
         description: `${modelToDelete?.displayName || 'Model'} has been deleted successfully.`,
       });
       setDeleteDialogOpen(false);
       setModelToDelete(null);
-      // Refetch status to update UI
+      // Invalidate AND explicitly refetch to ensure UI updates
-      queryClient.invalidateQueries({ queryKey: ['modelStatus'] });
+      // Using refetchType: 'all' ensures we refetch even if the query is stale
+      console.log('[Delete] Invalidating modelStatus query');
+      await queryClient.invalidateQueries({ 
+        queryKey: ['modelStatus'],
+        refetchType: 'all',
+      });
+      // Also explicitly refetch to guarantee fresh data
+      console.log('[Delete] Explicitly refetching modelStatus query');
+      await queryClient.refetchQueries({ queryKey: ['modelStatus'] });
+      console.log('[Delete] Query refetched');
     },
     onError: (error: Error) => {
+      console.log('[Delete] onError:', error);
       toast({
         title: 'Delete failed',
         description: error.message,
@@ -124,7 +169,7 @@ export function ModelManagement() {
                     <ModelItem
                       key={model.model_name}
                       model={model}
-                      onDownload={() => downloadMutation.mutate(model.model_name)}
+                      onDownload={() => handleDownload(model.model_name)}
                       onDelete={() => {
                         setModelToDelete({
                           name: model.model_name,
@@ -152,7 +197,7 @@ export function ModelManagement() {
                     <ModelItem
                       key={model.model_name}
                       model={model}
-                      onDownload={() => downloadMutation.mutate(model.model_name)}
+                      onDownload={() => handleDownload(model.model_name)}
                       onDelete={() => {
                         setModelToDelete({
                           name: model.model_name,
@@ -168,21 +213,6 @@ export function ModelManagement() {
               </div>
             </div>
 
-            {/* Progress indicators */}
-            <div className="pt-4 border-t">
-              <h3 className="text-sm font-semibold mb-3 text-muted-foreground">
-                Download Progress
-              </h3>
-              <div className="space-y-2">
-                {modelStatus.models.map((model) => (
-                  <ModelProgress
-                    key={model.model_name}
-                    modelName={model.model_name}
-                    displayName={model.display_name}
-                  />
-                ))}
-              </div>
-            </div>
           </div>
         ) : null}
       </CardContent>
@@ -235,16 +265,20 @@ interface ModelItemProps {
     model_name: string;
     display_name: string;
     downloaded: boolean;
+    downloading?: boolean;  // From server - true if download in progress
     size_mb?: number;
     loaded: boolean;
   };
   onDownload: () => void;
   onDelete: () => void;
-  isDownloading: boolean;
+  isDownloading: boolean;  // Local state - true if user just clicked download
   formatSize: (sizeMb?: number) => string;
 }
 
 function ModelItem({ model, onDownload, onDelete, isDownloading, formatSize }: ModelItemProps) {
+  // Use server's downloading state OR local state (for immediate feedback before server updates)
+  const showDownloading = model.downloading || isDownloading;
+  
   return (
     <div className="flex items-center justify-between p-3 border rounded-lg">
       <div className="flex-1">
@@ -255,20 +289,21 @@ function ModelItem({ model, onDownload, onDelete, isDownloading, formatSize }: M
               Loaded
             </Badge>
           )}
-          {model.downloaded && !model.loaded && (
+          {/* Only show Downloaded if actually downloaded AND not downloading */}
+          {model.downloaded && !model.loaded && !showDownloading && (
             <Badge variant="secondary" className="text-xs">
               Downloaded
             </Badge>
           )}
         </div>
-        {model.downloaded && model.size_mb && (
+        {model.downloaded && model.size_mb && !showDownloading && (
           <div className="text-xs text-muted-foreground mt-1">
             Size: {formatSize(model.size_mb)}
           </div>
         )}
       </div>
       <div className="flex items-center gap-2">
-        {model.downloaded ? (
+        {model.downloaded && !showDownloading ? (
           <div className="flex items-center gap-2">
             <div className="flex items-center gap-1 text-sm text-muted-foreground">
               <span>Ready</span>
@@ -283,19 +318,15 @@ function ModelItem({ model, onDownload, onDelete, isDownloading, formatSize }: M
               <Trash2 className="h-4 w-4" />
             </Button>
           </div>
-        ) : (
+        ) : showDownloading ? (
-          <Button size="sm" onClick={onDownload} disabled={isDownloading} variant="outline">
+          <Button size="sm" variant="outline" disabled>
-            {isDownloading ? (
-              <>
             <Loader2 className="h-4 w-4 mr-2 animate-spin" />
             Downloading...
-              </>
+          </Button>
         ) : (
-              <>
+          <Button size="sm" onClick={onDownload} variant="outline">
             <Download className="h-4 w-4 mr-2" />
             Download
-              </>
-            )}
           </Button>
         )}
       </div>

app/src/components/ServerSettings/ModelProgress.tsx

@@ -8,14 +8,23 @@ import { useServerStore } from '@/stores/serverStore';
 interface ModelProgressProps {
   modelName: string;
   displayName: string;
+  /** Only connect to SSE when actively downloading - prevents connection exhaustion */
+  isDownloading?: boolean;
 }
 
-export function ModelProgress({ modelName, displayName }: ModelProgressProps) {
+export function ModelProgress({ modelName, displayName, isDownloading = false }: ModelProgressProps) {
   const [progress, setProgress] = useState<ModelProgressType | null>(null);
   const serverUrl = useServerStore((state) => state.serverUrl);
 
   useEffect(() => {
-    if (!serverUrl) return;
+    // IMPORTANT: Only connect to SSE when this specific model is downloading
+    // Opening SSE connections for all models exhausts HTTP/1.1 connection limits (6 per origin)
+    // which causes other fetches (like the download trigger) to be queued/blocked
+    if (!serverUrl || !isDownloading) {
+      return;
+    }
+
+    console.log(`[ModelProgress] Connecting SSE for ${modelName}`);
 
     // Subscribe to progress updates via Server-Sent Events
     const eventSource = new EventSource(`${serverUrl}/models/progress/${modelName}`);
@@ -27,6 +36,7 @@ export function ModelProgress({ modelName, displayName }: ModelProgressProps) {
 
         // Close connection if complete or error
         if (data.status === 'complete' || data.status === 'error') {
+          console.log(`[ModelProgress] Download ${data.status} for ${modelName}, closing SSE`);
           eventSource.close();
         }
       } catch (error) {
@@ -35,14 +45,15 @@ export function ModelProgress({ modelName, displayName }: ModelProgressProps) {
     };
 
     eventSource.onerror = (error) => {
-      console.error('SSE error:', error);
+      console.error(`[ModelProgress] SSE error for ${modelName}:`, error);
       eventSource.close();
     };
 
     return () => {
+      console.log(`[ModelProgress] Cleanup - closing SSE for ${modelName}`);
       eventSource.close();
     };
-  }, [serverUrl, modelName]);
+  }, [serverUrl, modelName, isDownloading]);
 
   // Don't render if no progress or if complete/error and some time has passed
   if (

app/src/components/ServerSettings/UpdateStatus.tsx

@@ -13,9 +13,10 @@ export function UpdateStatus() {
   const [currentVersion, setCurrentVersion] = useState<string>('');
 
   useEffect(() => {
-    platform.metadata.getVersion()
+    platform.metadata
+      .getVersion()
       .then(setCurrentVersion)
-      .catch(() => setCurrentVersion('0.1.0'));
+      .catch(() => setCurrentVersion('Unknown'));
   }, [platform]);
 
   return (

app/src/hooks/useAutoUpdater.ts

@@ -1,4 +1,4 @@
-import { useCallback, useEffect, useState } from 'react';
+import { useCallback, useEffect, useRef, useState } from 'react';
 import { usePlatform } from '@/platform/PlatformContext';
 import type { UpdateStatus } from '@/platform/types';
 
@@ -7,9 +7,8 @@ export type { UpdateStatus };
 
 export function useAutoUpdater(checkOnMount = false) {
   const platform = usePlatform();
-  const [status, setStatus] = useState<UpdateStatus>(
+  const [status, setStatus] = useState<UpdateStatus>(platform.updater.getStatus());
-    platform.updater.getStatus(),
+  const hasCheckedRef = useRef(false);
-  );
 
   // Subscribe to updater status changes
   useEffect(() => {
@@ -17,25 +16,32 @@ export function useAutoUpdater(checkOnMount = false) {
       setStatus(newStatus);
     });
     return unsubscribe;
-  }, [platform]);
+    // Empty dependency array - platform is stable from context
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.updater.subscribe]);
 
   const checkForUpdates = useCallback(async () => {
     await platform.updater.checkForUpdates();
-  }, [platform]);
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.updater.checkForUpdates]);
 
   const downloadAndInstall = useCallback(async () => {
     await platform.updater.downloadAndInstall();
-  }, [platform]);
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.updater.downloadAndInstall]);
 
   const restartAndInstall = useCallback(async () => {
     await platform.updater.restartAndInstall();
-  }, [platform]);
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.updater.restartAndInstall]);
 
   useEffect(() => {
-    if (checkOnMount && platform.metadata.isTauri) {
+    if (checkOnMount && platform.metadata.isTauri && !hasCheckedRef.current) {
+      hasCheckedRef.current = true;
       checkForUpdates();
     }
-  }, [checkOnMount, checkForUpdates, platform.metadata.isTauri]);
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.metadata.isTauricheckOnMountcheckForUpdates]);
 
   return {
     status,

app/src/hooks/useAutoUpdater.tsx

@@ -0,0 +1,209 @@
+import { Download, RefreshCw } from 'lucide-react';
+import { useCallback, useEffect, useRef, useState } from 'react';
+import { Progress } from '@/components/ui/progress';
+import { ToastAction } from '@/components/ui/toast';
+import { useToast } from '@/components/ui/use-toast';
+import { usePlatform } from '@/platform/PlatformContext';
+import type { UpdateStatus } from '@/platform/types';
+
+// Re-export UpdateStatus for backwards compatibility
+export type { UpdateStatus };
+
+interface UseAutoUpdaterOptions {
+  checkOnMount?: boolean;
+  showToast?: boolean;
+}
+
+export function useAutoUpdater(options: boolean | UseAutoUpdaterOptions = false) {
+  // Support both old boolean API and new options object
+  const { checkOnMount, showToast } =
+    typeof options === 'boolean'
+      ? { checkOnMount: options, showToast: false }
+      : { checkOnMount: options.checkOnMount ?? false, showToast: options.showToast ?? false };
+
+  const platform = usePlatform();
+  const { toast } = useToast();
+  const [status, setStatus] = useState<UpdateStatus>(platform.updater.getStatus());
+  const hasCheckedRef = useRef(false);
+  const toastIdRef = useRef<string | null>(null);
+  const toastUpdateRef = useRef<
+    | ((props: {
+        title?: React.ReactNode;
+        description?: React.ReactNode;
+        duration?: number;
+        variant?: 'default' | 'destructive';
+        open?: boolean;
+        action?: React.ReactElement<typeof ToastAction>;
+      }) => void)
+    | null
+  >(null);
+
+  // Subscribe to updater status changes
+  useEffect(() => {
+    const unsubscribe = platform.updater.subscribe((newStatus) => {
+      setStatus(newStatus);
+    });
+    return unsubscribe;
+    // Empty dependency array - platform is stable from context
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.updater.subscribe]);
+
+  const checkForUpdates = useCallback(async () => {
+    await platform.updater.checkForUpdates();
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.updater.checkForUpdates]);
+
+  const downloadAndInstall = useCallback(async () => {
+    await platform.updater.downloadAndInstall();
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.updater.downloadAndInstall]);
+
+  const restartAndInstall = useCallback(async () => {
+    await platform.updater.restartAndInstall();
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.updater.restartAndInstall]);
+
+  // Check for updates on mount
+  useEffect(() => {
+    if (checkOnMount && platform.metadata.isTauri && !hasCheckedRef.current) {
+      hasCheckedRef.current = true;
+      checkForUpdates().catch((error) => {
+        console.error('Auto update check failed:', error);
+      });
+    }
+    // Empty dependency array - only run once on mount
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [platform.metadata.isTauricheckOnMountcheckForUpdates]);
+
+  // Show toast when update is available
+  useEffect(() => {
+    if (
+      !showToast ||
+      !status.available ||
+      status.downloading ||
+      status.readyToInstall ||
+      toastIdRef.current
+    ) {
+      return;
+    }
+
+    const handleUpdateNow = async () => {
+      await downloadAndInstall();
+    };
+
+    const toastResult = toast({
+      title: 'Update Available',
+      description: `Version ${status.version} is ready to download.`,
+      duration: Infinity,
+      action: (
+        <ToastAction altText="Update now" onClick={handleUpdateNow}>
+          Update Now
+        </ToastAction>
+      ),
+    });
+
+    toastIdRef.current = toastResult.id;
+    // Type assertion needed because update function has broader type than our ref
+    toastUpdateRef.current = toastResult.update as typeof toastUpdateRef.current;
+  }, [
+    showToast,
+    status.available,
+    status.downloading,
+    status.readyToInstall,
+    status.version,
+    downloadAndInstall,
+    toast,
+  ]);
+
+  // Update toast when downloading
+  useEffect(() => {
+    if (!showToast || !status.downloading || !toastIdRef.current || !toastUpdateRef.current) {
+      return;
+    }
+
+    const progressPercent = status.downloadProgress || 0;
+    const progressText =
+      status.downloadedBytes !== undefined &&
+      status.totalBytes !== undefined &&
+      status.totalBytes > 0
+        ? `${(status.downloadedBytes / 1024 / 1024).toFixed(1)} MB / ${(status.totalBytes / 1024 / 1024).toFixed(1)} MB`
+        : '';
+
+    toastUpdateRef.current({
+      title: (
+        <div className="flex items-center gap-2">
+          <Download className="h-4 w-4 animate-pulse" />
+          <span>Downloading Update</span>
+        </div>
+      ),
+      description: (
+        <div className="space-y-2">
+          <div className="text-sm">Version {status.version}</div>
+          {progressPercent > 0 && (
+            <>
+              <Progress value={progressPercent} className="h-2" />
+              {progressText && <div className="text-xs text-muted-foreground">{progressText}</div>}
+            </>
+          )}
+        </div>
+      ),
+      duration: Infinity,
+    });
+  }, [
+    showToast,
+    status.downloading,
+    status.downloadProgress,
+    status.downloadedBytes,
+    status.totalBytes,
+    status.version,
+  ]);
+
+  // Update toast when ready to install
+  useEffect(() => {
+    if (!showToast || !status.readyToInstall || !toastIdRef.current || !toastUpdateRef.current) {
+      return;
+    }
+
+    const handleRestartNow = async () => {
+      await restartAndInstall();
+    };
+
+    toastUpdateRef.current({
+      title: 'Update Ready',
+      description: `Version ${status.version} has been downloaded and is ready to install.`,
+      duration: Infinity,
+      action: (
+        <ToastAction altText="Restart now" onClick={handleRestartNow}>
+          <RefreshCw className="h-3 w-3 mr-1" />
+          Restart Now
+        </ToastAction>
+      ),
+    });
+  }, [showToast, status.readyToInstall, status.version, restartAndInstall]);
+
+  // Handle errors in toast
+  useEffect(() => {
+    if (!showToast || !status.error || !toastIdRef.current || !toastUpdateRef.current) {
+      return;
+    }
+
+    toastUpdateRef.current({
+      title: 'Update Failed',
+      description: status.error,
+      variant: 'destructive',
+      duration: 5000,
+    });
+
+    setTimeout(() => {
+      toastIdRef.current = null;
+      toastUpdateRef.current = null;
+    }, 5000);
+  }, [showToast, status.error]);
+
+  return {
+    status,
+    checkForUpdates,
+    downloadAndInstall,
+    restartAndInstall,
+  };
+}

app/src/lib/api/client.ts

@@ -310,10 +310,13 @@ class ApiClient {
   }
 
   async triggerModelDownload(modelName: string): Promise<{ message: string }> {
-    return this.request<{ message: string }>('/models/download', {
+    console.log('[API] triggerModelDownload called for:', modelName, 'at', new Date().toISOString());
+    const result = await this.request<{ message: string }>('/models/download', {
       method: 'POST',
       body: JSON.stringify({ model_name: modelName } as ModelDownloadRequest),
     });
+    console.log('[API] triggerModelDownload response:', result);
+    return result;
   }
 
   async deleteModel(modelName: string): Promise<{ message: string }> {

app/src/lib/api/models/ModelStatus.ts

@@ -9,6 +9,7 @@ export type ModelStatus = {
   model_name: string;
   display_name: string;
   downloaded: boolean;
+  downloading?: boolean;  // True if download is in progress
   size_mb?: number | null;
   loaded?: boolean;
 };

app/src/lib/api/types.ts

@@ -96,6 +96,7 @@ export interface ModelStatus {
   model_name: string;
   display_name: string;
   downloaded: boolean;
+  downloading: boolean;  // True if download is in progress
   size_mb?: number;
   loaded: boolean;
 }

app/src/lib/hooks/useModelDownloadToast.tsx

@@ -1,14 +1,16 @@
-import { useEffect, useRef } from 'react';
+import { CheckCircle2, Loader2, XCircle } from 'lucide-react';
-import { useToast } from '@/components/ui/use-toast';
+import { useCallback, useEffect, useRef } from 'react';
-import { useServerStore } from '@/stores/serverStore';
 import { Progress } from '@/components/ui/progress';
-import { Loader2, CheckCircle2, XCircle } from 'lucide-react';
+import { useToast } from '@/components/ui/use-toast';
 import type { ModelProgress } from '@/lib/api/types';
+import { useServerStore } from '@/stores/serverStore';
 
 interface UseModelDownloadToastOptions {
   modelName: string;
   displayName: string;
   enabled?: boolean;
+  onComplete?: () => void;
+  onError?: () => void;
 }
 
 /**
@@ -19,47 +21,64 @@ export function useModelDownloadToast({
   modelName,
   displayName,
   enabled = false,
+  onComplete,
+  onError,
 }: UseModelDownloadToastOptions) {
   const { toast } = useToast();
   const serverUrl = useServerStore((state) => state.serverUrl);
   const toastIdRef = useRef<string | null>(null);
-  const toastUpdateRef = useRef<
+  // biome-ignore lint: Using any for toast update ref to handle complex toast types
-    ((props: {
+  const toastUpdateRef = useRef<any>(null);
-      title?: React.ReactNode;
-      description?: React.ReactNode;
-      duration?: number;
-      variant?: 'default' | 'destructive';
-      open?: boolean;
-    }) => void) | null
-  >(null);
   const eventSourceRef = useRef<EventSource | null>(null);
 
-  const formatBytes = (bytes: number): string => {
+  const formatBytes = useCallback((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 / Math.pow(k, i)).toFixed(1)} ${sizes[i]}`;
+    return `${(bytes / k ** i).toFixed(1)} ${sizes[i]}`;
-  };
+  }, []);
 
   useEffect(() => {
+    console.log('[useModelDownloadToast] useEffect triggered', {
+      enabled,
+      serverUrl,
+      modelName,
+      displayName,
+    });
+
     if (!enabled || !serverUrl || !modelName) {
+      console.log('[useModelDownloadToast] Not enabled, skipping');
       return;
     }
 
+    console.log('[useModelDownloadToast] Creating toast and EventSource for:', modelName);
+
     // Create initial toast
     const toastResult = toast({
       title: displayName,
-      description: 'Starting download...',
+      description: (
+        <div className="flex items-center gap-2">
+          <Loader2 className="h-4 w-4 animate-spin" />
+          <span>Connecting to download...</span>
+        </div>
+      ),
       duration: Infinity, // Don't auto-dismiss, we'll handle it manually
     });
     toastIdRef.current = toastResult.id;
     toastUpdateRef.current = toastResult.update;
 
     // Subscribe to progress updates via Server-Sent Events
-    const eventSource = new EventSource(`${serverUrl}/models/progress/${modelName}`);
+    const eventSourceUrl = `${serverUrl}/models/progress/${modelName}`;
+    console.log('[useModelDownloadToast] Creating EventSource to:', eventSourceUrl);
+    const eventSource = new EventSource(eventSourceUrl);
+
+    eventSource.onopen = () => {
+      console.log('[useModelDownloadToast] EventSource connection opened for:', modelName);
+    };
 
     eventSource.onmessage = (event) => {
+      console.log('[useModelDownloadToast] Received SSE message:', event.data);
       try {
         const progress = JSON.parse(event.data) as ModelProgress;
 
@@ -86,7 +105,7 @@ export function useModelDownloadToast({
               break;
             case 'downloading':
               statusIcon = <Loader2 className="h-4 w-4 animate-spin" />;
-              statusText = progress.filename ? `Downloading ${progress.filename}...` : 'Downloading...';
+              statusText = progress.filename || 'Downloading...';
               break;
             case 'extracting':
               statusIcon = <Loader2 className="h-4 w-4 animate-spin" />;
@@ -117,21 +136,40 @@ export function useModelDownloadToast({
           });
 
           // Close connection and dismiss toast on completion or error
-          if (progress.status === 'complete' || progress.status === 'error') {
+          // Also treat progress >= 100% as complete
+          const isComplete = progress.status === 'complete' || progress.progress >= 100;
+          const isError = progress.status === 'error';
+
+          if (isComplete || isError) {
+            console.log('[useModelDownloadToast] Download finished:', {
+              isComplete,
+              isError,
+              progress: progress.progress,
+            });
             eventSource.close();
             eventSourceRef.current = null;
 
-            // Auto-dismiss on completion after delay
+            // Update toast to show completion state before callbacks
-            if (progress.status === 'complete') {
+            if (isComplete && toastUpdateRef.current) {
-              setTimeout(() => {
-                if (toastIdRef.current && toastUpdateRef.current) {
               toastUpdateRef.current({
-                    open: false,
+                title: (
+                  <div className="flex items-center gap-2">
+                    <CheckCircle2 className="h-4 w-4 text-green-500" />
+                    <span>{displayName}</span>
+                  </div>
+                ),
+                description: 'Download complete',
+                duration: 3000,
               });
-                  toastIdRef.current = null;
-                  toastUpdateRef.current = null;
             }
-              }, 5000);
+
+            // Call callbacks
+            if (isComplete && onComplete) {
+              console.log('[useModelDownloadToast] Download complete, calling onComplete callback');
+              onComplete();
+            } else if (isError && onError) {
+              console.log('[useModelDownloadToast] Download error, calling onError callback');
+              onError();
             }
           }
         }
@@ -141,7 +179,8 @@ export function useModelDownloadToast({
     };
 
     eventSource.onerror = (error) => {
-      console.error('SSE error:', error);
+      console.error('[useModelDownloadToast] SSE error for:', modelName, error);
+      console.log('[useModelDownloadToast] EventSource readyState:', eventSource.readyState);
       eventSource.close();
       eventSourceRef.current = null;
 
@@ -162,13 +201,14 @@ export function useModelDownloadToast({
 
     // Cleanup on unmount or when disabled
     return () => {
+      console.log('[useModelDownloadToast] Cleanup - closing EventSource for:', modelName);
       if (eventSourceRef.current) {
         eventSourceRef.current.close();
         eventSourceRef.current = null;
       }
       // Note: We don't dismiss the toast here as it might still be showing completion state
     };
-  }, [enabled, serverUrl, modelName, displayName, toast]);
+  }, [enabled, serverUrl, modelName, displayName, toast, formatBytes, onComplete, onError]);
 
   return {
     isTracking: enabled && eventSourceRef.current !== null,

backend/__init__.py

@@ -1,3 +1,3 @@
 # Backend package
 
-__version__ = "0.1.11"
+__version__ = "0.1.12"

backend/backends/mlx_backend.py

@@ -52,6 +52,47 @@ class MLXTTSBackend:
         
         return hf_model_id
     
+    def _is_model_cached(self, model_size: str) -> bool:
+        """
+        Check if the model is already cached locally AND fully downloaded.
+        
+        Args:
+            model_size: Model size to check
+            
+        Returns:
+            True if model is fully cached, False if missing or incomplete
+        """
+        try:
+            from huggingface_hub import constants as hf_constants
+            model_path = self._get_model_path(model_size)
+            repo_cache = Path(hf_constants.HF_HUB_CACHE) / ("models--" + model_path.replace("/", "--"))
+            
+            if not repo_cache.exists():
+                return False
+            
+            # Check for .incomplete files - if any exist, download is still in progress
+            blobs_dir = repo_cache / "blobs"
+            if blobs_dir.exists() and any(blobs_dir.glob("*.incomplete")):
+                print(f"[_is_model_cached] Found .incomplete files for {model_size}, treating as not cached")
+                return False
+            
+            # Check that actual model weight files exist in snapshots
+            snapshots_dir = repo_cache / "snapshots"
+            if snapshots_dir.exists():
+                has_weights = (
+                    any(snapshots_dir.rglob("*.safetensors")) or
+                    any(snapshots_dir.rglob("*.bin")) or
+                    any(snapshots_dir.rglob("*.npz"))
+                )
+                if not has_weights:
+                    print(f"[_is_model_cached] No model weights found for {model_size}, treating as not cached")
+                    return False
+            
+            return True
+        except Exception as e:
+            print(f"[_is_model_cached] Error checking cache for {model_size}: {e}")
+            return False
+    
     async def load_model_async(self, model_size: Optional[str] = None):
         """
         Lazy load the MLX TTS model.
@@ -79,46 +120,63 @@ class MLXTTSBackend:
     def _load_model_sync(self, model_size: str):
         """Synchronous model loading."""
         try:
-            from mlx_audio.tts import load
+            # Get model path BEFORE importing mlx_audio
-            
-            # Get model path
             model_path = self._get_model_path(model_size)
             
             # Set up progress tracking
             progress_manager = get_progress_manager()
+            task_manager = get_task_manager()
             model_name = f"qwen-tts-{model_size}"
             
-            # Start tracking download task
+            # Check if model is already cached
-            task_manager = get_task_manager()
+            is_cached = self._is_model_cached(model_size)
-            task_manager.start_download(model_name)
+            
+            # Set up progress callback
+            # If cached: filter out non-download progress
+            # If not cached: report all progress (we're actually downloading)
+            progress_callback = create_hf_progress_callback(model_name, progress_manager)
+            tracker = HFProgressTracker(progress_callback, filter_non_downloads=is_cached)
             
             print(f"Loading MLX TTS model {model_size}...")
             
-            # Initialize progress state
+            # Only track download progress if model is NOT cached
+            if not is_cached:
+                # Start tracking download task
+                task_manager.start_download(model_name)
+                
+                # Initialize progress state so SSE endpoint has initial data to send
+                # This provides immediate feedback while HuggingFace fetches metadata
                 progress_manager.update_progress(
                     model_name=model_name,
                     current=0,
-                total=1,
+                    total=0,  # Will be updated once actual total is known
-                filename="",
+                    filename="Connecting to HuggingFace...",
                     status="downloading",
                 )
             
-            # Set up progress callback
+            # IMPORTANT: Patch tqdm BEFORE importing mlx_audio
-            progress_callback = create_hf_progress_callback(model_name, progress_manager)
+            # Otherwise mlx_audio caches reference to original tqdm
-            tracker = HFProgressTracker(progress_callback)
+            tracker_context = tracker.patch_download()
+            tracker_context.__enter__()
+            
+            # Import mlx_audio AFTER patching tqdm
+            from mlx_audio.tts import load
             
-            # Use progress tracker during download
-            with tracker.patch_download():
             # Load MLX model (downloads automatically)
+            try:
                 self.model = load(model_path)
+            finally:
+                # Exit the patch context
+                tracker_context.__exit__(None, None, None)
+            
+            # Only mark download as complete if we were tracking it
+            if not is_cached:
+                progress_manager.mark_complete(model_name)
+                task_manager.complete_download(model_name)
             
             self._current_model_size = model_size
             self.model_size = model_size
             
-            # Mark as complete
-            progress_manager.mark_complete(model_name)
-            task_manager.complete_download(model_name)
-            
             print(f"MLX TTS model {model_size} loaded successfully")
             
         except ImportError as e:
@@ -332,6 +390,47 @@ class MLXSTTBackend:
         """Check if model is loaded."""
         return self.model is not None
     
+    def _is_model_cached(self, model_size: str) -> bool:
+        """
+        Check if the Whisper model is already cached locally AND fully downloaded.
+        
+        Args:
+            model_size: Model size to check
+            
+        Returns:
+            True if model is fully cached, False if missing or incomplete
+        """
+        try:
+            from huggingface_hub import constants as hf_constants
+            model_name = f"openai/whisper-{model_size}"
+            repo_cache = Path(hf_constants.HF_HUB_CACHE) / ("models--" + model_name.replace("/", "--"))
+            
+            if not repo_cache.exists():
+                return False
+            
+            # Check for .incomplete files - if any exist, download is still in progress
+            blobs_dir = repo_cache / "blobs"
+            if blobs_dir.exists() and any(blobs_dir.glob("*.incomplete")):
+                print(f"[_is_model_cached] Found .incomplete files for whisper-{model_size}, treating as not cached")
+                return False
+            
+            # Check that actual model weight files exist in snapshots
+            snapshots_dir = repo_cache / "snapshots"
+            if snapshots_dir.exists():
+                has_weights = (
+                    any(snapshots_dir.rglob("*.safetensors")) or
+                    any(snapshots_dir.rglob("*.bin")) or
+                    any(snapshots_dir.rglob("*.npz"))
+                )
+                if not has_weights:
+                    print(f"[_is_model_cached] No model weights found for whisper-{model_size}, treating as not cached")
+                    return False
+            
+            return True
+        except Exception as e:
+            print(f"[_is_model_cached] Error checking cache for whisper-{model_size}: {e}")
+            return False
+    
     async def load_model_async(self, model_size: Optional[str] = None):
         """
         Lazy load the MLX Whisper model.
@@ -354,56 +453,59 @@ class MLXSTTBackend:
     def _load_model_sync(self, model_size: str):
         """Synchronous model loading."""
         try:
-            # IMPORTANT: Set up progress tracking BEFORE importing mlx_audio
-            # This ensures tqdm is patched before any HuggingFace Hub imports
             progress_manager = get_progress_manager()
+            task_manager = get_task_manager()
             progress_model_name = f"whisper-{model_size}"
 
+            # Check if model is already cached
+            is_cached = self._is_model_cached(model_size)
+
             # Set up progress callback and tracker
+            # If cached: filter out non-download progress
+            # If not cached: report all progress (we're actually downloading)
             progress_callback = create_hf_progress_callback(progress_model_name, progress_manager)
-            tracker = HFProgressTracker(progress_callback)
+            tracker = HFProgressTracker(progress_callback, filter_non_downloads=is_cached)
 
             # Patch tqdm BEFORE importing mlx_audio
-            # This is critical because mlx_audio imports huggingface_hub which imports tqdm
-            print("[DEBUG] Starting tqdm patch BEFORE mlx_audio import")
             tracker_context = tracker.patch_download()
             tracker_context.__enter__()
-            print("[DEBUG] tqdm patched, now importing mlx_audio")
 
-            # NOW import mlx_audio - it will use our patched tqdm
+            # Import mlx_audio
             from mlx_audio.stt import load
 
             # MLX Whisper uses the standard OpenAI models
             model_name = f"openai/whisper-{model_size}"
 
-            # Start tracking download task
-            task_manager = get_task_manager()
-            task_manager.start_download(progress_model_name)
-
             print(f"Loading MLX Whisper model {model_size}...")
 
-            # Initialize progress state
+            # Only track download progress if model is NOT cached
+            if not is_cached:
+                # Start tracking download task
+                task_manager.start_download(progress_model_name)
+                
+                # Initialize progress state so SSE endpoint has initial data to send
                 progress_manager.update_progress(
                     model_name=progress_model_name,
                     current=0,
-                total=1,
+                    total=0,
-                filename="",
+                    filename="Connecting to HuggingFace...",
                     status="downloading",
                 )
 
-            # Load the model (tqdm is already patched from above)
+            # Load the model (tqdm is patched, but filters out non-download progress)
             try:
                 self.model = load(model_name)
             finally:
                 # Exit the patch context
                 tracker_context.__exit__(None, None, None)
             
-            self.model_size = model_size
+            # Only mark download as complete if we were tracking it
-            
+            if not is_cached:
-            # Mark as complete
                 progress_manager.mark_complete(progress_model_name)
                 task_manager.complete_download(progress_model_name)
             
+            self.model_size = model_size
+            
             print(f"MLX Whisper model {model_size} loaded successfully")
             
         except ImportError as e:

backend/backends/pytorch_backend.py

@@ -58,6 +58,46 @@ class PyTorchTTSBackend:
         
         return hf_model_map[model_size]
     
+    def _is_model_cached(self, model_size: str) -> bool:
+        """
+        Check if the model is already cached locally AND fully downloaded.
+        
+        Args:
+            model_size: Model size to check
+            
+        Returns:
+            True if model is fully cached, False if missing or incomplete
+        """
+        try:
+            from huggingface_hub import constants as hf_constants
+            model_path = self._get_model_path(model_size)
+            repo_cache = Path(hf_constants.HF_HUB_CACHE) / ("models--" + model_path.replace("/", "--"))
+            
+            if not repo_cache.exists():
+                return False
+            
+            # Check for .incomplete files - if any exist, download is still in progress
+            blobs_dir = repo_cache / "blobs"
+            if blobs_dir.exists() and any(blobs_dir.glob("*.incomplete")):
+                print(f"[_is_model_cached] Found .incomplete files for {model_size}, treating as not cached")
+                return False
+            
+            # Check that actual model weight files exist in snapshots
+            snapshots_dir = repo_cache / "snapshots"
+            if snapshots_dir.exists():
+                has_weights = (
+                    any(snapshots_dir.rglob("*.safetensors")) or
+                    any(snapshots_dir.rglob("*.bin"))
+                )
+                if not has_weights:
+                    print(f"[_is_model_cached] No model weights found for {model_size}, treating as not cached")
+                    return False
+            
+            return True
+        except Exception as e:
+            print(f"[_is_model_cached] Error checking cache for {model_size}: {e}")
+            return False
+    
     async def load_model_async(self, model_size: Optional[str] = None):
         """
         Lazy load the TTS model with automatic downloading from HuggingFace Hub.
@@ -85,20 +125,24 @@ class PyTorchTTSBackend:
     def _load_model_sync(self, model_size: str):
         """Synchronous model loading."""
         try:
-            # IMPORTANT: Set up progress tracking BEFORE importing qwen_tts
-            # This ensures tqdm is patched before any HuggingFace Hub imports
             progress_manager = get_progress_manager()
+            task_manager = get_task_manager()
             model_name = f"qwen-tts-{model_size}"
 
+            # Check if model is already cached
+            is_cached = self._is_model_cached(model_size)
+
             # Set up progress callback and tracker
+            # If cached: filter out non-download progress (like "Segment 1/1" during generation)
+            # If not cached: report all progress (we're actually downloading)
             progress_callback = create_hf_progress_callback(model_name, progress_manager)
-            tracker = HFProgressTracker(progress_callback)
+            tracker = HFProgressTracker(progress_callback, filter_non_downloads=is_cached)
 
             # Patch tqdm BEFORE importing qwen_tts
             tracker_context = tracker.patch_download()
             tracker_context.__enter__()
 
-            # NOW import qwen_tts - it will use our patched tqdm
+            # Import qwen_tts
             from qwen_tts import Qwen3TTSModel
 
             # Get model path (local or HuggingFace Hub ID)
@@ -106,20 +150,21 @@ class PyTorchTTSBackend:
 
             print(f"Loading TTS model {model_size} on {self.device}...")
 
+            # Only track download progress if model is NOT cached
+            if not is_cached:
                 # Start tracking download task
-            task_manager = get_task_manager()
                 task_manager.start_download(model_name)
 
-            # Initialize progress state to show download has started
+                # Initialize progress state so SSE endpoint has initial data to send
                 progress_manager.update_progress(
                     model_name=model_name,
                     current=0,
-                total=1,  # Set to 1 initially, will be updated by callback
+                    total=0,  # Will be updated once actual total is known
-                filename="",
+                    filename="Connecting to HuggingFace...",
                     status="downloading",
                 )
 
-            # Load the model (tqdm is already patched from above)
+            # Load the model (tqdm is patched, but filters out non-download progress)
             try:
                 self.model = Qwen3TTSModel.from_pretrained(
                     model_path,
@@ -130,7 +175,8 @@ class PyTorchTTSBackend:
                 # Exit the patch context
                 tracker_context.__exit__(None, None, None)
             
-            # Mark as complete
+            # Only mark download as complete if we were tracking it
+            if not is_cached:
                 progress_manager.mark_complete(model_name)
                 task_manager.complete_download(model_name)
             
@@ -321,6 +367,46 @@ class PyTorchSTTBackend:
         """Check if model is loaded."""
         return self.model is not None
     
+    def _is_model_cached(self, model_size: str) -> bool:
+        """
+        Check if the Whisper model is already cached locally AND fully downloaded.
+        
+        Args:
+            model_size: Model size to check
+            
+        Returns:
+            True if model is fully cached, False if missing or incomplete
+        """
+        try:
+            from huggingface_hub import constants as hf_constants
+            model_name = f"openai/whisper-{model_size}"
+            repo_cache = Path(hf_constants.HF_HUB_CACHE) / ("models--" + model_name.replace("/", "--"))
+            
+            if not repo_cache.exists():
+                return False
+            
+            # Check for .incomplete files - if any exist, download is still in progress
+            blobs_dir = repo_cache / "blobs"
+            if blobs_dir.exists() and any(blobs_dir.glob("*.incomplete")):
+                print(f"[_is_model_cached] Found .incomplete files for whisper-{model_size}, treating as not cached")
+                return False
+            
+            # Check that actual model weight files exist in snapshots
+            snapshots_dir = repo_cache / "snapshots"
+            if snapshots_dir.exists():
+                has_weights = (
+                    any(snapshots_dir.rglob("*.safetensors")) or
+                    any(snapshots_dir.rglob("*.bin"))
+                )
+                if not has_weights:
+                    print(f"[_is_model_cached] No model weights found for whisper-{model_size}, treating as not cached")
+                    return False
+            
+            return True
+        except Exception as e:
+            print(f"[_is_model_cached] Error checking cache for whisper-{model_size}: {e}")
+            return False
+    
     async def load_model_async(self, model_size: Optional[str] = None):
         """
         Lazy load the Whisper model.
@@ -349,14 +435,18 @@ class PyTorchSTTBackend:
         """Synchronous model loading."""
         print(f"[DEBUG] _load_model_sync called for Whisper {model_size}")
         try:
-            # IMPORTANT: Set up progress tracking BEFORE importing transformers
-            # This ensures tqdm is patched before any HuggingFace Hub imports
             progress_manager = get_progress_manager()
+            task_manager = get_task_manager()
             progress_model_name = f"whisper-{model_size}"
 
+            # Check if model is already cached
+            is_cached = self._is_model_cached(model_size)
+
             # Set up progress callback and tracker
+            # If cached: filter out non-download progress
+            # If not cached: report all progress (we're actually downloading)
             progress_callback = create_hf_progress_callback(progress_model_name, progress_manager)
-            tracker = HFProgressTracker(progress_callback)
+            tracker = HFProgressTracker(progress_callback, filter_non_downloads=is_cached)
 
             # Patch tqdm BEFORE importing transformers
             print("[DEBUG] Starting tqdm patch BEFORE transformers import")
@@ -364,31 +454,29 @@ class PyTorchSTTBackend:
             tracker_context.__enter__()
             print("[DEBUG] tqdm patched, now importing transformers")
 
-            # NOW import transformers - it will use our patched tqdm
+            # Import transformers
             from transformers import WhisperProcessor, WhisperForConditionalGeneration
 
             model_name = f"openai/whisper-{model_size}"
             print(f"[DEBUG] Model name: {model_name}")
 
-            # Start tracking download task
-            task_manager = get_task_manager()
-            task_manager.start_download(progress_model_name)
-            print(f"[DEBUG] Task manager started download")
-
             print(f"Loading Whisper model {model_size} on {self.device}...")
 
-            # Initialize progress state to show download has started
+            # Only track download progress if model is NOT cached
-            print(f"[DEBUG] Calling update_progress...")
+            if not is_cached:
+                # Start tracking download task
+                task_manager.start_download(progress_model_name)
+
+                # Initialize progress state so SSE endpoint has initial data to send
                 progress_manager.update_progress(
                     model_name=progress_model_name,
                     current=0,
-                total=1,  # Set to 1 initially, will be updated by callback
+                    total=0,  # Will be updated once actual total is known
-                filename="",
+                    filename="Connecting to HuggingFace...",
                     status="downloading",
                 )
-            print(f"[DEBUG] update_progress called, listeners: {len(progress_manager._listeners.get(progress_model_name, []))}")
 
-            # Load models (tqdm is already patched from above)
+            # Load models (tqdm is patched, but filters out non-download progress)
             try:
                 self.processor = WhisperProcessor.from_pretrained(model_name)
                 self.model = WhisperForConditionalGeneration.from_pretrained(model_name)
@@ -396,13 +484,14 @@ class PyTorchSTTBackend:
                 # Exit the patch context
                 tracker_context.__exit__(None, None, None)
             
-            self.model.to(self.device)
+            # Only mark download as complete if we were tracking it
-            self.model_size = model_size
+            if not is_cached:
-            
-            # Mark as complete
                 progress_manager.mark_complete(progress_model_name)
                 task_manager.complete_download(progress_model_name)
             
+            self.model.to(self.device)
+            self.model_size = model_size
+            
             print(f"Whisper model {model_size} loaded successfully")
             
         except Exception as e:

backend/build_binary.py

@@ -5,6 +5,7 @@ PyInstaller build script for creating standalone Python server binary.
 import PyInstaller.__main__
 import os
 import platform
+import sys
 from pathlib import Path
 
 
@@ -13,15 +14,27 @@ def is_apple_silicon():
     return platform.system() == "Darwin" and platform.machine() == "arm64"
 
 
-def build_server():
+def build_server(variant="cpu"):
-    """Build Python server as standalone binary."""
+    """Build Python server as standalone binary.
+
+    Args:
+        variant: 'cpu' for CPU-only build (~500MB) or 'cuda' for CUDA build (~3GB)
+    """
     backend_dir = Path(__file__).parent
 
+    if variant not in ['cpu', 'cuda']:
+        raise ValueError(f"Invalid variant: {variant}. Must be 'cpu' or 'cuda'")
+
+    # Set binary name based on variant
+    binary_name = f'voicebox-server-{variant}' if variant == 'cuda' else 'voicebox-server'
+
+    print(f"Building {variant.upper()} variant: {binary_name}")
+
     # PyInstaller arguments
     args = [
         'server.py',  # Use server.py as entry point instead of main.py
         '--onefile',
-        '--name', 'voicebox-server',
+        '--name', binary_name,
     ]
 
     # Add local qwen_tts path if specified (for editable installs)
@@ -101,8 +114,13 @@ def build_server():
     # Run PyInstaller
     PyInstaller.__main__.run(args)
 
-    print(f"Binary built in {backend_dir / 'dist' / 'voicebox-server'}")
+    print(f"\n{'='*60}")
+    print(f"Build complete: {variant.upper()} variant")
+    print(f"Binary: {backend_dir / 'dist' / binary_name}")
+    print(f"{'='*60}\n")
 
 
 if __name__ == '__main__':
-    build_server()
+    # Accept variant as command line argument
+    variant = sys.argv[1] if len(sys.argv) > 1 else 'cpu'
+    build_server(variant)

backend/build_both.bat

@@ -0,0 +1,30 @@
+@echo off
+REM Build both CPU and CUDA server binaries for Windows
+
+echo ============================================================
+echo Building BOTH server binaries (CPU + CUDA)
+echo This will take a while...
+echo ============================================================
+
+call build_cpu.bat
+if errorlevel 1 (
+    echo CPU build failed!
+    exit /b 1
+)
+
+echo.
+echo.
+
+call build_cuda.bat
+if errorlevel 1 (
+    echo CUDA build failed!
+    exit /b 1
+)
+
+echo.
+echo ============================================================
+echo Both binaries built successfully!
+echo ============================================================
+echo CPU binary:  dist\voicebox-server.exe (~500MB)
+echo CUDA binary: dist\voicebox-server-cuda.exe (~3GB)
+echo ============================================================

backend/build_cpu.bat

@@ -0,0 +1,28 @@
+@echo off
+REM Build CPU-only server binary for Windows
+REM This creates a ~500MB binary without CUDA support
+
+echo ============================================================
+echo Building CPU-only server binary
+echo ============================================================
+
+echo.
+echo Step 1: Installing CPU-only PyTorch...
+pip uninstall -y torch torchvision torchaudio
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
+
+echo.
+echo Step 2: Building binary with PyInstaller...
+python build_binary.py cpu
+
+echo.
+echo Step 3: Restoring CUDA PyTorch for development...
+pip uninstall -y torch torchvision torchaudio
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
+
+echo.
+echo ============================================================
+echo CPU binary built successfully!
+echo Location: dist\voicebox-server.exe
+echo Size: ~500MB
+echo ============================================================

backend/build_cpu.sh

@@ -0,0 +1,30 @@
+#!/bin/bash
+# Build CPU-only server binary
+# This creates a ~500MB binary without CUDA support
+
+set -e
+
+echo "============================================================"
+echo "Building CPU-only server binary"
+echo "============================================================"
+
+echo ""
+echo "Step 1: Installing CPU-only PyTorch..."
+pip uninstall -y torch torchvision torchaudio || true
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
+
+echo ""
+echo "Step 2: Building binary with PyInstaller..."
+python build_binary.py cpu
+
+echo ""
+echo "Step 3: Restoring CUDA PyTorch for development..."
+pip uninstall -y torch torchvision torchaudio || true
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
+
+echo ""
+echo "============================================================"
+echo "CPU binary built successfully!"
+echo "Location: dist/voicebox-server"
+echo "Size: ~500MB"
+echo "============================================================"

backend/build_cuda.bat

@@ -0,0 +1,22 @@
+@echo off
+REM Build CUDA server binary for Windows
+REM This creates a ~3GB binary with CUDA support
+
+echo ============================================================
+echo Building CUDA server binary
+echo ============================================================
+
+echo.
+echo Step 1: Ensuring CUDA PyTorch is installed...
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121 --upgrade
+
+echo.
+echo Step 2: Building binary with PyInstaller...
+python build_binary.py cuda
+
+echo.
+echo ============================================================
+echo CUDA binary built successfully!
+echo Location: dist\voicebox-server-cuda.exe
+echo Size: ~3GB
+echo ============================================================

backend/main.py

@@ -1156,11 +1156,14 @@ async def get_model_progress(model_name: str):
 @app.get("/models/status", response_model=models.ModelStatusListResponse)
 async def get_model_status():
     """Get status of all available models."""
-    from huggingface_hub import hf_hub_download, constants as hf_constants
+    from huggingface_hub import constants as hf_constants
     from pathlib import Path
-    import os
     
     backend_type = get_backend_type()
+    task_manager = get_task_manager()
+    
+    # Get set of currently downloading model names
+    active_download_names = {task.model_name for task in task_manager.get_active_downloads()}
     
     # Try to import scan_cache_dir (might not be available in older versions)
     try:
@@ -1189,10 +1192,11 @@ async def get_model_status():
     if backend_type == "mlx":
         tts_1_7b_id = "mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16"
         tts_0_6b_id = "mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16"  # Fallback to 1.7B
-        whisper_base_id = "mlx-community/whisper-base"
+        # MLX backend uses openai/whisper-* models, not mlx-community
-        whisper_small_id = "mlx-community/whisper-small"
+        whisper_base_id = "openai/whisper-base"
-        whisper_medium_id = "mlx-community/whisper-medium"
+        whisper_small_id = "openai/whisper-small"
-        whisper_large_id = "mlx-community/whisper-large"
+        whisper_medium_id = "openai/whisper-medium"
+        whisper_large_id = "openai/whisper-large"
     else:
         tts_1_7b_id = "Qwen/Qwen3-TTS-12Hz-1.7B-Base"
         tts_0_6b_id = "Qwen/Qwen3-TTS-12Hz-0.6B-Base"
@@ -1246,6 +1250,13 @@ async def get_model_status():
         },
     ]
     
+    # Build a mapping of model_name -> hf_repo_id so we can check if shared repos are downloading
+    model_to_repo = {cfg["model_name"]: cfg["hf_repo_id"] for cfg in model_configs}
+    
+    # Get the set of hf_repo_ids that are currently being downloaded
+    # This handles the case where multiple models share the same repo (e.g., 0.6B and 1.7B on MLX)
+    active_download_repos = {model_to_repo.get(name) for name in active_download_names if name in model_to_repo}
+    
     # Get HuggingFace cache info (if available)
     cache_info = None
     if use_scan_cache:
@@ -1268,6 +1279,30 @@ async def get_model_status():
                 repo_id = config["hf_repo_id"]
                 for repo in cache_info.repos:
                     if repo.repo_id == repo_id:
+                        # Check if actual model weight files exist (not just config files)
+                        # scan_cache_dir only shows completed files, so check if any are model weights
+                        has_model_weights = False
+                        for rev in repo.revisions:
+                            for f in rev.files:
+                                fname = f.file_name.lower()
+                                if fname.endswith(('.safetensors', '.bin', '.pt', '.pth', '.npz')):
+                                    has_model_weights = True
+                                    break
+                            if has_model_weights:
+                                break
+                        
+                        # Also check for .incomplete files in blobs directory (downloads in progress)
+                        has_incomplete = False
+                        try:
+                            cache_dir = hf_constants.HF_HUB_CACHE
+                            blobs_dir = Path(cache_dir) / ("models--" + repo_id.replace("/", "--")) / "blobs"
+                            if blobs_dir.exists():
+                                has_incomplete = any(blobs_dir.glob("*.incomplete"))
+                        except Exception:
+                            pass
+                        
+                        # Only mark as downloaded if we have model weights AND no incomplete files
+                        if has_model_weights and not has_incomplete:
                             downloaded = True
                             # Calculate size from cache info
                             try:
@@ -1284,42 +1319,40 @@ async def get_model_status():
                     repo_cache = Path(cache_dir) / ("models--" + config["hf_repo_id"].replace("/", "--"))
                     
                     if repo_cache.exists():
-                        # Check for model files (bin, safetensors, or other common model files)
+                        # Check for .incomplete files - if any exist, download is still in progress
-                        # MLX models may use .npz or .safetensors
+                        blobs_dir = repo_cache / "blobs"
+                        has_incomplete = blobs_dir.exists() and any(blobs_dir.glob("*.incomplete"))
+                        
+                        if not has_incomplete:
+                            # Check for actual model weight files (not just index files)
+                            # in the snapshots directory (symlinks to completed blobs)
+                            snapshots_dir = repo_cache / "snapshots"
+                            has_model_files = False
+                            if snapshots_dir.exists():
                                 has_model_files = (
-                            any(repo_cache.rglob("*.bin")) or
+                                    any(snapshots_dir.rglob("*.bin")) or
-                            any(repo_cache.rglob("*.safetensors")) or
+                                    any(snapshots_dir.rglob("*.safetensors")) or
-                            any(repo_cache.rglob("*.pt")) or
+                                    any(snapshots_dir.rglob("*.pt")) or
-                            any(repo_cache.rglob("*.pth")) or
+                                    any(snapshots_dir.rglob("*.pth")) or
-                            any(repo_cache.rglob("*.npz")) or
+                                    any(snapshots_dir.rglob("*.npz"))
-                            any(repo_cache.rglob("model.safetensors.index.json")) or
-                            any(repo_cache.rglob("pytorch_model.bin.index.json"))
                                 )
                             
                             if has_model_files:
                                 downloaded = True
-                            # Calculate size
+                                # Calculate size (exclude .incomplete files)
                                 try:
-                                total_size = sum(f.stat().st_size for f in repo_cache.rglob("*") if f.is_file())
+                                    total_size = sum(
+                                        f.stat().st_size for f in repo_cache.rglob("*") 
+                                        if f.is_file() and not f.name.endswith('.incomplete')
+                                    )
                                     size_mb = total_size / (1024 * 1024)
                                 except Exception:
                                     pass
                 except Exception:
                     pass
             
-            # Method 3: Try to check if model can be loaded locally (last resort)
+            # Method 3 removed - checking for config.json is too lenient
-            if not downloaded:
+            # Methods 1 and 2 properly verify that model weight files exist
-                try:
-                    # Try to download with local_files_only=True to check if cached
-                    hf_hub_download(
-                        repo_id=config["hf_repo_id"],
-                        filename="config.json",  # Try a common file
-                        local_files_only=True,
-                    )
-                    downloaded = True
-                except Exception:
-                    # File not found locally, model not downloaded
-                    pass
             
             # Check if loaded in memory
             try:
@@ -1327,10 +1360,19 @@ async def get_model_status():
             except Exception:
                 loaded = False
             
+            # Check if this model (or its shared repo) is currently being downloaded
+            is_downloading = config["hf_repo_id"] in active_download_repos
+            
+            # If downloading, don't report as downloaded (partial files exist)
+            if is_downloading:
+                downloaded = False
+                size_mb = None  # Don't show partial size during download
+            
             statuses.append(models.ModelStatus(
                 model_name=config["model_name"],
                 display_name=config["display_name"],
                 downloaded=downloaded,
+                downloading=is_downloading,
                 size_mb=size_mb,
                 loaded=loaded,
             ))
@@ -1341,10 +1383,14 @@ async def get_model_status():
             except Exception:
                 loaded = False
             
+            # Check if this model (or its shared repo) is currently being downloaded
+            is_downloading = config["hf_repo_id"] in active_download_repos
+            
             statuses.append(models.ModelStatus(
                 model_name=config["model_name"],
                 display_name=config["display_name"],
                 downloaded=False,  # Assume not downloaded if check failed
+                downloading=is_downloading,
                 size_mb=None,
                 loaded=loaded,
             ))
@@ -1358,6 +1404,7 @@ async def trigger_model_download(request: models.ModelDownloadRequest):
     import asyncio
     
     task_manager = get_task_manager()
+    progress_manager = get_progress_manager()
     
     model_configs = {
         "qwen-tts-1.7B": {
@@ -1406,6 +1453,18 @@ async def trigger_model_download(request: models.ModelDownloadRequest):
     # Start tracking download
     task_manager.start_download(request.model_name)
     
+    # Initialize progress state so SSE endpoint has initial data to send.
+    # This fixes a race condition where the frontend connects to SSE before
+    # any progress callbacks have fired (especially for large models like Qwen
+    # where huggingface_hub takes time to fetch metadata for all files).
+    progress_manager.update_progress(
+        model_name=request.model_name,
+        current=0,
+        total=0,  # Will be updated once actual total is known
+        filename="Connecting to HuggingFace...",
+        status="downloading",
+    )
+
     # Start download in background task (don't await)
     asyncio.create_task(download_in_background())
 

backend/models.py

@@ -134,6 +134,7 @@ class ModelStatus(BaseModel):
     model_name: str
     display_name: str
     downloaded: bool
+    downloading: bool = False  # True if download is in progress
     size_mb: Optional[float] = None
     loaded: bool = False
 

backend/test_cuda_compression.py

@@ -0,0 +1,137 @@
+#!/usr/bin/env python3
+"""
+Test CUDA binary compression to verify it fits under GitHub's 2GB release asset limit.
+
+Usage:
+    python test_cuda_compression.py [path/to/voicebox-server-cuda.exe]
+
+If no path provided, looks for the binary in ./dist/
+"""
+
+import os
+import sys
+import subprocess
+from pathlib import Path
+
+
+def format_size(bytes_size):
+    """Format bytes into human-readable size."""
+    for unit in ['B', 'KB', 'MB', 'GB']:
+        if bytes_size < 1024.0:
+            return f"{bytes_size:.2f} {unit}"
+        bytes_size /= 1024.0
+    return f"{bytes_size:.2f} TB"
+
+
+def get_file_size(filepath):
+    """Get file size in bytes."""
+    return os.path.getsize(filepath)
+
+
+def compress_with_7z(input_file, output_file):
+    """Compress file using 7z with maximum compression."""
+    print(f"\nCompressing with 7z (maximum compression)...")
+    print(f"This may take several minutes for a ~2.5GB file...\n")
+
+    cmd = [
+        '7z', 'a',
+        '-t7z',           # 7z format
+        '-m0=lzma2',      # LZMA2 compression
+        '-mx=9',          # Maximum compression
+        '-mfb=64',        # Fast bytes
+        '-md=32m',        # Dictionary size
+        '-ms=on',         # Solid archive
+        output_file,
+        input_file
+    ]
+
+    try:
+        subprocess.run(cmd, check=True, capture_output=True, text=True)
+        return True
+    except subprocess.CalledProcessError as e:
+        print(f"Error during compression: {e}")
+        print(f"stderr: {e.stderr}")
+        return False
+    except FileNotFoundError:
+        print("ERROR: 7z not found. Please install 7-Zip:")
+        print("  Windows: https://www.7-zip.org/download.html")
+        print("  macOS: brew install p7zip")
+        print("  Linux: apt-get install p7zip-full")
+        return False
+
+
+def main():
+    # Find CUDA binary
+    if len(sys.argv) > 1:
+        cuda_binary = Path(sys.argv[1])
+    else:
+        # Look in dist directory
+        dist_dir = Path(__file__).parent / 'dist'
+        candidates = list(dist_dir.glob('voicebox-server-cuda*.exe'))
+
+        if not candidates:
+            print("ERROR: CUDA binary not found in ./dist/")
+            print("Please provide the path as an argument:")
+            print("  python test_cuda_compression.py path/to/voicebox-server-cuda.exe")
+            sys.exit(1)
+
+        cuda_binary = candidates[0]
+
+    if not cuda_binary.exists():
+        print(f"ERROR: File not found: {cuda_binary}")
+        sys.exit(1)
+
+    print("=" * 70)
+    print("CUDA Binary Compression Test")
+    print("=" * 70)
+
+    # Get original size
+    original_size = get_file_size(cuda_binary)
+    print(f"\nOriginal file: {cuda_binary.name}")
+    print(f"Original size: {format_size(original_size)} ({original_size:,} bytes)")
+
+    # Check if already over 2GB
+    github_limit = 2 * 1024 * 1024 * 1024  # 2GB in bytes
+    print(f"GitHub limit:  {format_size(github_limit)} ({github_limit:,} bytes)")
+
+    if original_size > github_limit:
+        print(f"\n[WARNING] Original file exceeds GitHub limit by {format_size(original_size - github_limit)}")
+    else:
+        print(f"\n[OK] Original file is under GitHub limit")
+
+    # Compress
+    output_file = cuda_binary.parent / f"{cuda_binary.stem}.7z"
+    if output_file.exists():
+        print(f"\nRemoving existing compressed file: {output_file.name}")
+        output_file.unlink()
+
+    success = compress_with_7z(cuda_binary, output_file)
+
+    if not success:
+        sys.exit(1)
+
+    # Check compressed size
+    compressed_size = get_file_size(output_file)
+    compression_ratio = (1 - compressed_size / original_size) * 100
+
+    print("\n" + "=" * 70)
+    print("Compression Results")
+    print("=" * 70)
+    print(f"\nCompressed file: {output_file.name}")
+    print(f"Compressed size: {format_size(compressed_size)} ({compressed_size:,} bytes)")
+    print(f"Compression ratio: {compression_ratio:.1f}%")
+    print(f"Space saved: {format_size(original_size - compressed_size)}")
+
+    if compressed_size <= github_limit:
+        print(f"\n[SUCCESS] Compressed file fits under GitHub's 2GB limit!")
+        print(f"   Margin: {format_size(github_limit - compressed_size)} remaining")
+    else:
+        print(f"\n[FAILED] Compressed file still exceeds GitHub limit")
+        print(f"   Over by: {format_size(compressed_size - github_limit)}")
+        print(f"\n   Alternative: Host on external storage (S3, Azure Blob, etc.)")
+
+    print("\n" + "=" * 70)
+
+
+if __name__ == '__main__':
+    main()

backend/test_r2_upload.sh

@@ -0,0 +1,86 @@
+#!/bin/bash
+# Test R2 upload locally before running in CI
+
+set -e
+
+echo "============================================================"
+echo "Cloudflare R2 Upload Test"
+echo "============================================================"
+
+# Check for required environment variables
+if [ -z "$AWS_ACCESS_KEY_ID" ] || [ -z "$AWS_SECRET_ACCESS_KEY" ] || [ -z "$R2_ENDPOINT" ]; then
+    echo "ERROR: Missing required environment variables"
+    echo ""
+    echo "Please set:"
+    echo "  export AWS_ACCESS_KEY_ID='your-r2-access-key-id'"
+    echo "  export AWS_SECRET_ACCESS_KEY='your-r2-secret-access-key'"
+    echo "  export R2_ENDPOINT='https://your-account-id.r2.cloudflarestorage.com'"
+    echo ""
+    exit 1
+fi
+
+# Check for AWS CLI
+if ! command -v aws &> /dev/null; then
+    echo "Installing AWS CLI..."
+    pip install awscli
+fi
+
+# Find CUDA binary
+CUDA_BINARY=$(ls dist/voicebox-server-cuda*.exe 2>/dev/null | head -1)
+
+if [ -z "$CUDA_BINARY" ]; then
+    echo "ERROR: CUDA binary not found in dist/"
+    echo "Run: bash build_cuda.bat"
+    exit 1
+fi
+
+echo ""
+echo "Found CUDA binary: $CUDA_BINARY"
+echo "Size: $(du -h "$CUDA_BINARY" | cut -f1)"
+echo ""
+
+# Test version
+VERSION="v0.1.12-test"
+PLATFORM="x86_64-pc-windows-msvc"
+FILENAME="voicebox-server-cuda-${PLATFORM}.exe"
+
+echo "Test upload configuration:"
+echo "  Version: $VERSION"
+echo "  Platform: $PLATFORM"
+echo "  Endpoint: $R2_ENDPOINT"
+echo "  Bucket: voicebox"
+echo "  Path: cuda/$VERSION/$FILENAME"
+echo ""
+
+read -p "Proceed with upload? (y/n) " -n 1 -r
+echo
+if [[ ! $REPLY =~ ^[Yy]$ ]]; then
+    echo "Aborted."
+    exit 1
+fi
+
+echo ""
+echo "Uploading to R2..."
+
+aws s3 cp "$CUDA_BINARY" \
+    "s3://voicebox/cuda/${VERSION}/${FILENAME}" \
+    --endpoint-url "$R2_ENDPOINT" \
+    --acl public-read
+
+if [ $? -eq 0 ]; then
+    echo ""
+    echo "============================================================"
+    echo "Upload successful!"
+    echo "============================================================"
+    echo ""
+    echo "Download URL:"
+    echo "https://downloads.voicebox.sh/cuda/${VERSION}/${FILENAME}"
+    echo ""
+    echo "Test with:"
+    echo "curl -I https://downloads.voicebox.sh/cuda/${VERSION}/${FILENAME}"
+    echo ""
+else
+    echo ""
+    echo "Upload failed!"
+    exit 1
+fi

backend/tests/README.md

@@ -0,0 +1,58 @@
+# Backend Tests
+
+Manual test scripts for debugging and validating backend functionality.
+
+## Test Files
+
+### `test_generation_progress.py`
+Tests TTS generation with SSE progress monitoring to identify UX issues where users see download progress even when the model is already cached.
+
+**Usage:**
+```bash
+cd backend
+python tests/test_generation_progress.py
+```
+
+**Prerequisites:**
+- Server must be running (`python main.py`)
+- At least one voice profile must exist
+
+### `test_real_download.py`
+Tests real model download with SSE progress monitoring.
+
+**Usage:**
+```bash
+cd backend
+# Delete cache first to force fresh download
+rm -rf ~/.cache/huggingface/hub/models--openai--whisper-base
+python tests/test_real_download.py
+```
+
+**Prerequisites:**
+- Server must be running (`python main.py`)
+
+### `test_progress.py`
+Unit tests for ProgressManager and HFProgressTracker functionality.
+
+**Usage:**
+```bash
+cd backend
+python tests/test_progress.py
+```
+
+### `test_check_progress_state.py`
+Debugging script to inspect the internal state of ProgressManager and TaskManager.
+
+**Usage:**
+```bash
+cd backend
+python tests/test_check_progress_state.py
+```
+
+## Notes
+
+These are manual test scripts, not automated unit tests. They're designed for:
+- Debugging progress tracking issues
+- Validating SSE event streams
+- Monitoring real-time download behavior
+- Inspecting internal state during development

backend/tests/__init__.py

@@ -0,0 +1,6 @@
+"""
+Test suite for Voicebox backend.
+
+This directory contains manual test scripts for debugging and validating
+progress tracking, model downloads, and generation functionality.
+"""

backend/tests/test_generation_download.py

@@ -0,0 +1,321 @@
+"""
+Test TTS generation with SSE progress monitoring.
+This test captures the exact SSE events triggered during generation
+to identify UX issues where users see download progress even when
+the model is already cached.
+"""
+
+import asyncio
+import json
+import httpx
+from typing import List, Dict, Optional
+from datetime import datetime
+
+
+async def monitor_sse_stream(model_name: str, timeout: int = 120):
+    """Monitor SSE stream for a model during generation."""
+    events: List[Dict] = []
+    url = f"http://localhost:8000/models/progress/{model_name}"
+
+    print(f"[{_timestamp()}] Connecting to SSE endpoint: {url}")
+
+    try:
+        async with httpx.AsyncClient(timeout=timeout) as client:
+            async with client.stream("GET", url) as response:
+                print(f"[{_timestamp()}] SSE connected, status: {response.status_code}")
+
+                if response.status_code != 200:
+                    print(f"[{_timestamp()}] Error: SSE endpoint returned {response.status_code}")
+                    return events
+
+                async for line in response.aiter_lines():
+                    if not line:
+                        continue
+
+                    timestamp = _timestamp()
+
+                    if line.startswith("data: "):
+                        try:
+                            data = json.loads(line[6:])
+                            print(f"[{timestamp}] → SSE Event: {data['status']:12} {data.get('progress', 0):6.1f}% {data.get('filename', '')}")
+                            events.append({
+                                **data,
+                                "_timestamp": timestamp
+                            })
+
+                            # Stop if complete or error
+                            if data.get("status") in ("complete", "error"):
+                                print(f"[{timestamp}] → Model {data['status']}!")
+                                break
+
+                        except json.JSONDecodeError as e:
+                            print(f"[{timestamp}] Error parsing JSON: {e}")
+                            print(f"  Line was: {line}")
+
+                    elif line.startswith(": heartbeat"):
+                        print(f"[{timestamp}] ♥ heartbeat")
+
+    except asyncio.TimeoutError:
+        print(f"[{_timestamp()}] SSE monitoring timed out")
+    except Exception as e:
+        print(f"[{_timestamp()}] SSE error: {e}")
+
+    return events
+
+
+async def trigger_generation(profile_id: str, text: str, model_size: str = "1.7B"):
+    """Trigger TTS generation via the API."""
+    url = "http://localhost:8000/generate"
+
+    print(f"\n[{_timestamp()}] Triggering generation...")
+    print(f"  Profile: {profile_id}")
+    print(f"  Text: {text[:50]}...")
+    print(f"  Model: {model_size}")
+
+    try:
+        async with httpx.AsyncClient(timeout=120) as client:
+            response = await client.post(url, json={
+                "profile_id": profile_id,
+                "text": text,
+                "language": "en",
+                "model_size": model_size,
+            })
+
+            print(f"[{_timestamp()}] Response: {response.status_code}")
+
+            if response.status_code == 200:
+                result = response.json()
+                print(f"[{_timestamp()}] ✓ Generation successful!")
+                print(f"  Generation ID: {result.get('id')}")
+                print(f"  Duration: {result.get('duration', 0):.2f}s")
+                return True, result
+            elif response.status_code == 202:
+                # Model is being downloaded
+                result = response.json()
+                print(f"[{_timestamp()}] → Model download in progress")
+                print(f"  Detail: {result}")
+                return False, result
+            else:
+                print(f"[{_timestamp()}] ✗ Error: {response.text}")
+                return False, None
+
+    except Exception as e:
+        print(f"[{_timestamp()}] ✗ Exception: {e}")
+        return False, None
+
+
+async def get_first_profile():
+    """Get the first available voice profile."""
+    url = "http://localhost:8000/profiles"
+
+    try:
+        async with httpx.AsyncClient(timeout=10) as client:
+            response = await client.get(url)
+            if response.status_code == 200:
+                profiles = response.json()
+                if profiles:
+                    return profiles[0]["id"]
+    except Exception as e:
+        print(f"Error getting profiles: {e}")
+
+    return None
+
+
+async def check_server():
+    """Check if the server is running."""
+    try:
+        async with httpx.AsyncClient(timeout=5) as client:
+            response = await client.get("http://localhost:8000/health")
+            return response.status_code == 200
+    except Exception as e:
+        print(f"Server not running: {e}")
+        return False
+
+
+def _timestamp():
+    """Get current timestamp for logging."""
+    return datetime.now().strftime("%H:%M:%S.%f")[:-3]
+
+
+async def test_generation_with_cached_model():
+    """
+    Test Case 1: Generation when model is already cached.
+    
+    This should NOT show any download progress events.
+    If it does, that's the UX bug we're trying to fix.
+    """
+    print("\n" + "=" * 80)
+    print("TEST CASE 1: Generation with Cached Model")
+    print("=" * 80)
+    print("Expected: No download progress events (or minimal/instant completion)")
+    print("Actual UX Issue: Users see 'started' and 'finished' events even for cached models")
+    print("=" * 80)
+
+    model_size = "1.7B"
+    model_name = f"qwen-tts-{model_size}"
+
+    # Get a profile
+    profile_id = await get_first_profile()
+    if not profile_id:
+        print("✗ No voice profiles found. Please create a profile first.")
+        return False
+
+    print(f"\nUsing profile: {profile_id}")
+
+    # Start SSE monitor BEFORE triggering generation
+    monitor_task = asyncio.create_task(monitor_sse_stream(model_name, timeout=30))
+
+    # Wait for SSE to connect
+    await asyncio.sleep(1)
+
+    # Trigger generation
+    test_text = "Hello, this is a test of the voice generation system."
+    success, result = await trigger_generation(profile_id, test_text, model_size)
+
+    if not success and result and result.get("downloading"):
+        print("\n⚠ Model is being downloaded. Waiting for download to complete...")
+        # Wait for SSE monitor to capture download events
+        events = await monitor_task
+        return events
+
+    # Wait a bit more to catch any progress events
+    await asyncio.sleep(3)
+
+    # Cancel SSE monitor
+    monitor_task.cancel()
+    try:
+        events = await monitor_task
+    except asyncio.CancelledError:
+        events = []
+
+    return events
+
+
+async def test_generation_with_fresh_download():
+    """
+    Test Case 2: Generation when model needs to be downloaded.
+    
+    This SHOULD show download progress events.
+    """
+    print("\n" + "=" * 80)
+    print("TEST CASE 2: Generation with Model Download")
+    print("=" * 80)
+    print("Expected: Download progress events from 0% to 100%")
+    print("=" * 80)
+
+    # Use a different model size to force download
+    model_size = "0.6B"  # Smaller model for faster testing
+    model_name = f"qwen-tts-{model_size}"
+
+    # Get a profile
+    profile_id = await get_first_profile()
+    if not profile_id:
+        print("✗ No voice profiles found. Please create a profile first.")
+        return False
+
+    print(f"\nUsing profile: {profile_id}")
+    print("Note: This will download the model if not cached")
+
+    # Start SSE monitor BEFORE triggering generation
+    monitor_task = asyncio.create_task(monitor_sse_stream(model_name, timeout=300))
+
+    # Wait for SSE to connect
+    await asyncio.sleep(1)
+
+    # Trigger generation
+    test_text = "This should trigger a model download if the model is not cached."
+    success, result = await trigger_generation(profile_id, test_text, model_size)
+
+    if not success and result and result.get("downloading"):
+        print("\n→ Model download initiated. Monitoring progress...")
+        # Wait for download to complete
+        events = await monitor_task
+
+        # Try generation again
+        print(f"\n[{_timestamp()}] Retrying generation after download...")
+        await asyncio.sleep(2)
+        success, result = await trigger_generation(profile_id, test_text, model_size)
+
+        if success:
+            print("✓ Generation successful after download")
+
+        return events
+
+    # If model was already cached
+    await asyncio.sleep(3)
+    monitor_task.cancel()
+    try:
+        events = await monitor_task
+    except asyncio.CancelledError:
+        events = []
+
+    return events
+
+
+async def main():
+    print("=" * 80)
+    print("TTS Generation Progress Test")
+    print("=" * 80)
+    print("Purpose: Capture exact SSE events during generation to identify UX issues")
+    print("=" * 80)
+
+    # Check if server is running
+    print(f"\n[{_timestamp()}] Checking if server is running...")
+    if not await check_server():
+        print("✗ Server is not running on http://localhost:8000")
+        print("\nPlease start the server first:")
+        print("  cd backend && python main.py")
+        return False
+
+    print("✓ Server is running")
+
+    # Test Case 1: Cached model
+    print("\n" + "🧪 " * 20)
+    events_cached = await test_generation_with_cached_model()
+
+    # Results for Test Case 1
+    print("\n" + "=" * 80)
+    print("TEST CASE 1 RESULTS: Generation with Cached Model")
+    print("=" * 80)
+
+    if not events_cached:
+        print("✓ GOOD: No SSE progress events received")
+        print("  This is the expected behavior for a cached model.")
+    else:
+        print(f"⚠ ISSUE FOUND: Received {len(events_cached)} SSE events:")
+        print("\nEvent Timeline:")
+        for i, event in enumerate(events_cached, 1):
+            timestamp = event.pop("_timestamp", "??:??:??.???")
+            print(f"  {i}. [{timestamp}] {event}")
+
+        print("\n⚠ This explains the UX issue!")
+        print("  Users see progress events even when the model is already cached,")
+        print("  making them think the model is downloading again.")
+
+    # Test Case 2: Fresh download (optional, commented out by default)
+    # Uncomment if you want to test download progress
+    # print("\n" + "🧪 " * 20)
+    # events_download = await test_generation_with_fresh_download()
+    #
+    # print("\n" + "=" * 80)
+    # print("TEST CASE 2 RESULTS: Generation with Model Download")
+    # print("=" * 80)
+    #
+    # if not events_download:
+    #     print("ℹ Model was already cached, no download occurred")
+    # else:
+    #     print(f"✓ Received {len(events_download)} download progress events")
+    #     print("\nDownload Timeline:")
+    #     for i, event in enumerate(events_download, 1):
+    #         timestamp = event.pop("_timestamp", "??:??:??.???")
+    #         print(f"  {i}. [{timestamp}] {event}")
+
+    print("\n" + "=" * 80)
+    print("Test Complete!")
+    print("=" * 80)
+
+    return True
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

backend/tests/test_progress.py

@@ -0,0 +1,313 @@
+"""
+Test script to debug model download progress tracking.
+"""
+
+import asyncio
+import json
+import time
+from typing import List, Dict
+import logging
+
+# Set up logging to see what's happening
+logging.basicConfig(
+    level=logging.DEBUG,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+)
+
+from utils.progress import ProgressManager, get_progress_manager
+from utils.hf_progress import HFProgressTracker, create_hf_progress_callback
+
+
+def test_progress_manager_basic():
+    """Test 1: Basic ProgressManager functionality."""
+    print("\n" + "=" * 60)
+    print("Test 1: ProgressManager Basic Operations")
+    print("=" * 60)
+
+    pm = ProgressManager()
+
+    # Test update_progress
+    pm.update_progress(
+        model_name="test-model",
+        current=50,
+        total=100,
+        filename="test.bin",
+        status="downloading"
+    )
+
+    # Test get_progress
+    progress = pm.get_progress("test-model")
+    print(f"✓ Progress stored: {progress}")
+    assert progress is not None
+    assert progress["progress"] == 50.0
+    assert progress["filename"] == "test.bin"
+    assert progress["status"] == "downloading"
+
+    # Test mark_complete
+    pm.mark_complete("test-model")
+    progress = pm.get_progress("test-model")
+    print(f"✓ Marked complete: {progress}")
+    assert progress["status"] == "complete"
+    assert progress["progress"] == 100.0
+
+    print("✓ Test 1 PASSED\n")
+    return True
+
+
+async def test_progress_manager_sse():
+    """Test 2: ProgressManager SSE streaming."""
+    print("\n" + "=" * 60)
+    print("Test 2: ProgressManager SSE Streaming")
+    print("=" * 60)
+
+    pm = ProgressManager()
+    collected_events: List[Dict] = []
+
+    # Simulate SSE client
+    async def sse_client():
+        """Simulates a frontend SSE connection."""
+        print("  SSE client: Subscribing to test-model-sse...")
+        async for event in pm.subscribe("test-model-sse"):
+            # Parse SSE event
+            if event.startswith("data: "):
+                data = json.loads(event[6:])
+                print(f"  SSE client: Received event: {data['status']} - {data.get('progress', 0):.1f}%")
+                collected_events.append(data)
+
+                # Stop when complete
+                if data.get("status") in ("complete", "error"):
+                    break
+            elif event.startswith(": heartbeat"):
+                print("  SSE client: Received heartbeat")
+
+    # Simulate download progress updates (from backend thread)
+    async def simulate_download():
+        """Simulates backend sending progress updates."""
+        print("  Backend: Starting simulated download...")
+        await asyncio.sleep(0.2)  # Let SSE client subscribe first
+
+        # Send progress updates
+        for i in range(0, 101, 20):
+            print(f"  Backend: Updating progress to {i}%")
+            pm.update_progress(
+                model_name="test-model-sse",
+                current=i,
+                total=100,
+                filename=f"file_{i}.bin",
+                status="downloading" if i < 100 else "downloading"
+            )
+            await asyncio.sleep(0.1)
+
+        # Mark complete
+        print("  Backend: Marking download complete")
+        pm.mark_complete("test-model-sse")
+
+    # Run SSE client and download simulation concurrently
+    await asyncio.gather(
+        sse_client(),
+        simulate_download()
+    )
+
+    # Verify we got events
+    print(f"\n  Collected {len(collected_events)} events")
+    assert len(collected_events) > 0, "Should have received at least one event"
+    assert collected_events[-1]["status"] == "complete", "Last event should be 'complete'"
+
+    print("✓ Test 2 PASSED\n")
+    return True
+
+
+def test_hf_progress_tracker():
+    """Test 3: HFProgressTracker tqdm patching."""
+    print("\n" + "=" * 60)
+    print("Test 3: HFProgressTracker tqdm Patching")
+    print("=" * 60)
+
+    captured_progress: List[tuple] = []
+
+    def progress_callback(downloaded: int, total: int, filename: str):
+        """Capture progress updates."""
+        captured_progress.append((downloaded, total, filename))
+        print(f"  Progress callback: {downloaded}/{total} bytes ({filename})")
+
+    tracker = HFProgressTracker(progress_callback)
+
+    # Simulate a download with tqdm
+    with tracker.patch_download():
+        try:
+            from tqdm import tqdm
+
+            # Simulate downloading a file
+            print("  Simulating download with tqdm...")
+            total_size = 1000
+            with tqdm(total=total_size, desc="model.bin", unit="B", unit_scale=True) as pbar:
+                for chunk in range(0, total_size, 100):
+                    pbar.update(100)
+                    time.sleep(0.01)
+
+            print(f"  Captured {len(captured_progress)} progress updates")
+            assert len(captured_progress) > 0, "Should have captured progress updates"
+
+            # Verify progress increases
+            last_downloaded = 0
+            for downloaded, total, filename in captured_progress:
+                assert downloaded >= last_downloaded, "Downloaded bytes should increase"
+                assert total == total_size, "Total should be consistent"
+                last_downloaded = downloaded
+
+            print("✓ Test 3 PASSED\n")
+            return True
+
+        except ImportError:
+            print("✗ tqdm not available, skipping test\n")
+            return None
+
+
+async def test_full_integration():
+    """Test 4: Full integration test."""
+    print("\n" + "=" * 60)
+    print("Test 4: Full Integration (ProgressManager + HFProgressTracker)")
+    print("=" * 60)
+
+    pm = get_progress_manager()
+    collected_events: List[Dict] = []
+
+    # SSE client
+    async def sse_client():
+        print("  SSE client: Subscribing...")
+        async for event in pm.subscribe("integration-test"):
+            if event.startswith("data: "):
+                data = json.loads(event[6:])
+                print(f"  SSE client: {data['status']} - {data.get('progress', 0):.1f}% - {data.get('filename', '')}")
+                collected_events.append(data)
+                if data.get("status") in ("complete", "error"):
+                    break
+
+    # Simulate backend download with HFProgressTracker
+    async def simulate_real_download():
+        await asyncio.sleep(0.2)  # Let SSE subscribe
+
+        print("  Backend: Starting download with HFProgressTracker...")
+
+        # Set up tracking (like the real backend does)
+        progress_callback = create_hf_progress_callback("integration-test", pm)
+        tracker = HFProgressTracker(progress_callback)
+
+        # Initialize progress
+        pm.update_progress(
+            model_name="integration-test",
+            current=0,
+            total=1,
+            filename="",
+            status="downloading"
+        )
+
+        # Simulate download with tqdm patching
+        with tracker.patch_download():
+            try:
+                from tqdm import tqdm
+
+                # Simulate multi-file download (like HuggingFace does)
+                files = [
+                    ("model.safetensors", 5000),
+                    ("config.json", 1000),
+                    ("tokenizer.json", 500),
+                ]
+
+                for filename, size in files:
+                    print(f"  Backend: Downloading {filename}...")
+                    with tqdm(total=size, desc=filename, unit="B") as pbar:
+                        for chunk in range(0, size, 500):
+                            chunk_size = min(500, size - chunk)
+                            pbar.update(chunk_size)
+                            await asyncio.sleep(0.05)
+
+                # Mark complete
+                print("  Backend: Download complete")
+                pm.mark_complete("integration-test")
+
+            except ImportError:
+                print("  ✗ tqdm not available")
+                pm.mark_error("integration-test", "tqdm not available")
+
+    # Run both
+    await asyncio.gather(
+        sse_client(),
+        simulate_real_download()
+    )
+
+    # Verify
+    print(f"\n  Collected {len(collected_events)} events")
+    if len(collected_events) > 0:
+        print(f"  First event: {collected_events[0]}")
+        print(f"  Last event: {collected_events[-1]}")
+        assert collected_events[-1]["status"] == "complete", "Should end with 'complete'"
+        print("✓ Test 4 PASSED\n")
+        return True
+    else:
+        print("✗ Test 4 FAILED - No events received\n")
+        return False
+
+
+async def main():
+    """Run all tests."""
+    print("\n" + "=" * 60)
+    print("Voicebox Progress Tracking Test Suite")
+    print("=" * 60)
+
+    results = []
+
+    # Test 1: Basic operations
+    try:
+        results.append(("Basic Operations", test_progress_manager_basic()))
+    except Exception as e:
+        print(f"✗ Test 1 FAILED: {e}\n")
+        results.append(("Basic Operations", False))
+
+    # Test 2: SSE streaming
+    try:
+        results.append(("SSE Streaming", await test_progress_manager_sse()))
+    except Exception as e:
+        print(f"✗ Test 2 FAILED: {e}\n")
+        results.append(("SSE Streaming", False))
+
+    # Test 3: tqdm patching
+    try:
+        results.append(("tqdm Patching", test_hf_progress_tracker()))
+    except Exception as e:
+        print(f"✗ Test 3 FAILED: {e}\n")
+        results.append(("tqdm Patching", False))
+
+    # Test 4: Full integration
+    try:
+        results.append(("Full Integration", await test_full_integration()))
+    except Exception as e:
+        print(f"✗ Test 4 FAILED: {e}\n")
+        results.append(("Full Integration", False))
+
+    # Summary
+    print("\n" + "=" * 60)
+    print("Test Results Summary")
+    print("=" * 60)
+
+    for name, result in results:
+        status = "✓ PASS" if result else ("⊘ SKIP" if result is None else "✗ FAIL")
+        print(f"  {status:8} {name}")
+
+    passed = sum(1 for _, r in results if r is True)
+    failed = sum(1 for _, r in results if r is False)
+    skipped = sum(1 for _, r in results if r is None)
+
+    print()
+    print(f"  Total: {len(results)} tests")
+    print(f"  Passed: {passed}")
+    print(f"  Failed: {failed}")
+    print(f"  Skipped: {skipped}")
+    print("=" * 60 + "\n")
+
+    return failed == 0
+
+
+if __name__ == "__main__":
+    success = asyncio.run(main())
+    exit(0 if success else 1)

backend/tests/test_qwen_download.py

@@ -0,0 +1,317 @@
+"""
+Test Qwen TTS model download with SSE progress monitoring.
+
+This specifically tests the MLX TTS backend download progress tracking,
+which requires tqdm to be patched BEFORE mlx_audio is imported.
+
+Usage:
+    cd backend && python -m tests.test_qwen_download
+
+Prerequisites:
+    - Server must be running: cd backend && python main.py
+    - Delete model first for fresh download test:
+      curl -X DELETE http://localhost:8000/models/qwen-tts-0.6B
+"""
+
+import asyncio
+import json
+import httpx
+import time
+from typing import List, Dict, Optional
+
+
+async def monitor_sse_stream(model_name: str, timeout: int = 600) -> List[Dict]:
+    """
+    Monitor SSE stream for a model download.
+    
+    Args:
+        model_name: Name of the model to monitor
+        timeout: Maximum time to wait for download (seconds)
+        
+    Returns:
+        List of SSE events received
+    """
+    events: List[Dict] = []
+    url = f"http://localhost:8000/models/progress/{model_name}"
+    last_progress = -1
+    
+    print(f"\n📡 Connecting to SSE endpoint: {url}")
+
+    try:
+        async with httpx.AsyncClient(timeout=timeout) as client:
+            async with client.stream("GET", url) as response:
+                print(f"   SSE connected, status: {response.status_code}")
+
+                if response.status_code != 200:
+                    print(f"   ❌ Error: SSE endpoint returned {response.status_code}")
+                    return events
+
+                async for line in response.aiter_lines():
+                    if not line:
+                        continue
+
+                    if line.startswith("data: "):
+                        try:
+                            data = json.loads(line[6:])
+                            events.append(data)
+                            
+                            # Print progress (only when it changes significantly)
+                            progress = data.get('progress', 0)
+                            status = data.get('status', 'unknown')
+                            filename = data.get('filename', '')
+                            current = data.get('current', 0)
+                            total = data.get('total', 0)
+                            
+                            # Print every 5% change or status change
+                            if abs(progress - last_progress) >= 5 or status in ('complete', 'error'):
+                                current_mb = current / (1024 * 1024)
+                                total_mb = total / (1024 * 1024)
+                                print(f"   📊 {status:12} {progress:6.1f}% ({current_mb:.1f}MB / {total_mb:.1f}MB) {filename[:50]}")
+                                last_progress = progress
+
+                            # Stop if complete or error
+                            if status in ("complete", "error"):
+                                if status == "complete":
+                                    print(f"   ✅ Download complete!")
+                                else:
+                                    print(f"   ❌ Download error: {data.get('error', 'unknown')}")
+                                break
+
+                        except json.JSONDecodeError as e:
+                            print(f"   ⚠️  Error parsing JSON: {e}")
+
+                    elif line.startswith(": heartbeat"):
+                        # Heartbeat every 1 second, don't spam
+                        pass
+
+    except asyncio.CancelledError:
+        print("   ⏹️  SSE monitor cancelled")
+    except Exception as e:
+        print(f"   ❌ SSE error: {e}")
+
+    return events
+
+
+async def trigger_download(model_name: str) -> bool:
+    """Trigger a model download via the API."""
+    url = "http://localhost:8000/models/download"
+
+    print(f"\n🚀 Triggering download for: {model_name}")
+
+    try:
+        async with httpx.AsyncClient(timeout=30) as client:
+            response = await client.post(url, json={"model_name": model_name})
+            result = response.json()
+            print(f"   Response: {response.status_code} - {result}")
+            return response.status_code == 200
+    except Exception as e:
+        print(f"   ❌ Error triggering download: {e}")
+        return False
+
+
+async def delete_model(model_name: str) -> bool:
+    """Delete a model from cache."""
+    url = f"http://localhost:8000/models/{model_name}"
+
+    print(f"\n🗑️  Deleting model: {model_name}")
+
+    try:
+        async with httpx.AsyncClient(timeout=30) as client:
+            response = await client.delete(url)
+            if response.status_code == 200:
+                print(f"   ✅ Model deleted")
+                return True
+            elif response.status_code == 404:
+                print(f"   ℹ️  Model not found (already deleted)")
+                return True
+            else:
+                print(f"   ⚠️  Delete response: {response.status_code} - {response.text}")
+                return False
+    except Exception as e:
+        print(f"   ❌ Error deleting model: {e}")
+        return False
+
+
+async def check_model_status(model_name: str) -> Optional[Dict]:
+    """Check the status of a model."""
+    try:
+        async with httpx.AsyncClient(timeout=10) as client:
+            response = await client.get("http://localhost:8000/models/status")
+            if response.status_code == 200:
+                data = response.json()
+                for model in data.get("models", []):
+                    if model["model_name"] == model_name:
+                        return model
+    except Exception as e:
+        print(f"   ⚠️  Error checking model status: {e}")
+    return None
+
+
+async def check_server() -> bool:
+    """Check if the server is running."""
+    try:
+        async with httpx.AsyncClient(timeout=5) as client:
+            response = await client.get("http://localhost:8000/health")
+            return response.status_code == 200
+    except Exception:
+        return False
+
+
+async def main():
+    print("=" * 70)
+    print("🧪 Qwen TTS Model Download Progress Test")
+    print("=" * 70)
+    print("\nThis test verifies that MLX TTS download progress tracking works.")
+    print("It specifically tests the tqdm patching for mlx_audio.tts imports.")
+
+    # Check if server is running
+    print("\n📡 Checking if server is running...")
+    if not await check_server():
+        print("   ❌ Server is not running on http://localhost:8000")
+        print("\n   Please start the server first:")
+        print("   cd backend && python main.py")
+        return False
+
+    print("   ✅ Server is running")
+
+    # Test model
+    model_name = "qwen-tts-0.6B"  # Note: 0.6B currently maps to 1.7B on MLX
+    
+    # Check current status
+    print(f"\n📊 Checking status of {model_name}...")
+    status = await check_model_status(model_name)
+    if status:
+        print(f"   Downloaded: {status.get('downloaded', False)}")
+        print(f"   Downloading: {status.get('downloading', False)}")
+        print(f"   Loaded: {status.get('loaded', False)}")
+        if status.get('size_mb'):
+            print(f"   Size: {status['size_mb']:.1f} MB")
+    else:
+        print("   ⚠️  Could not get model status")
+
+    # Ask if user wants to delete first
+    print("\n" + "-" * 70)
+    if status and status.get('downloaded'):
+        print("⚠️  Model is already downloaded. Delete it for a fresh download test?")
+        print("   [y] Yes, delete and download fresh")
+        print("   [n] No, just test SSE connection")
+        print("   [q] Quit")
+        
+        choice = input("\nChoice [y/n/q]: ").strip().lower()
+        
+        if choice == 'q':
+            print("Exiting...")
+            return True
+        
+        if choice == 'y':
+            if not await delete_model(model_name):
+                print("Failed to delete model. Continue anyway? [y/n]")
+                if input().strip().lower() != 'y':
+                    return False
+    else:
+        print("Model not downloaded. Will perform fresh download test.")
+        input("Press Enter to continue...")
+
+    # Run the test
+    print("\n" + "=" * 70)
+    print("🏃 Starting Download Test")
+    print("=" * 70)
+
+    async def run_test():
+        # Start SSE monitor in background FIRST
+        monitor_task = asyncio.create_task(monitor_sse_stream(model_name, timeout=600))
+
+        # Wait for SSE to connect
+        await asyncio.sleep(1)
+
+        # Trigger download
+        success = await trigger_download(model_name)
+
+        if not success:
+            print("   ❌ Failed to trigger download")
+            monitor_task.cancel()
+            try:
+                await monitor_task
+            except asyncio.CancelledError:
+                pass
+            return []
+
+        # Wait for SSE monitor to complete
+        print("\n⏳ Waiting for download to complete (this may take several minutes)...")
+        events = await monitor_task
+
+        return events
+
+    start_time = time.time()
+    events = await run_test()
+    elapsed = time.time() - start_time
+
+    # Results
+    print("\n" + "=" * 70)
+    print("📋 Test Results")
+    print("=" * 70)
+
+    print(f"\n⏱️  Elapsed time: {elapsed:.1f} seconds")
+    print(f"📨 Total SSE events received: {len(events)}")
+
+    if not events:
+        print("\n❌ FAILED - No SSE events received!")
+        print("\nPossible causes:")
+        print("  1. SSE endpoint not working")
+        print("  2. tqdm not patched before mlx_audio import")
+        print("  3. Progress callbacks not firing")
+        print("  4. Model already fully downloaded")
+        print("\nDebug steps:")
+        print("  1. Check server logs for [DEBUG] messages")
+        print("  2. Look for 'tqdm patched' before 'mlx_audio.tts import'")
+        print(f"  3. Delete model: curl -X DELETE http://localhost:8000/models/{model_name}")
+        return False
+
+    # Analyze events
+    first_event = events[0]
+    last_event = events[-1]
+    
+    print(f"\n📊 First event:")
+    print(f"   Status: {first_event.get('status')}")
+    print(f"   Progress: {first_event.get('progress', 0):.1f}%")
+    
+    print(f"\n📊 Last event:")
+    print(f"   Status: {last_event.get('status')}")
+    print(f"   Progress: {last_event.get('progress', 0):.1f}%")
+
+    # Check for expected behaviors
+    has_progress_updates = len(events) > 2
+    has_increasing_progress = False
+    has_complete = any(e.get('status') == 'complete' for e in events)
+    has_100_percent = any(e.get('progress', 0) >= 100 for e in events)
+    
+    # Check if progress increased over time
+    if len(events) >= 2:
+        progress_values = [e.get('progress', 0) for e in events]
+        has_increasing_progress = progress_values[-1] > progress_values[0]
+
+    print("\n📋 Checks:")
+    print(f"   {'✅' if has_progress_updates else '❌'} Multiple progress updates received ({len(events)} events)")
+    print(f"   {'✅' if has_increasing_progress else '❌'} Progress increased over time")
+    print(f"   {'✅' if has_100_percent else '❌'} Reached 100% progress")
+    print(f"   {'✅' if has_complete else '❌'} Received 'complete' status")
+
+    # Overall result
+    success = has_progress_updates and has_complete
+    
+    if success:
+        print("\n" + "=" * 70)
+        print("✅ TEST PASSED - Qwen TTS download progress tracking works!")
+        print("=" * 70)
+    else:
+        print("\n" + "=" * 70)
+        print("❌ TEST FAILED - Progress tracking has issues")
+        print("=" * 70)
+        print("\nCheck the server logs for debug output.")
+
+    return success
+
+
+if __name__ == "__main__":
+    result = asyncio.run(main())
+    exit(0 if result else 1)

backend/tests/test_whisper_download.py

@@ -0,0 +1,178 @@
+"""
+Test real model download with SSE progress monitoring.
+"""
+
+import asyncio
+import json
+import httpx
+import time
+from typing import List, Dict
+
+async def monitor_sse_stream(model_name: str, timeout: int = 300):
+    """Monitor SSE stream for a model download."""
+    events: List[Dict] = []
+    url = f"http://localhost:8000/models/progress/{model_name}"
+
+    print(f"Connecting to SSE endpoint: {url}")
+
+    async with httpx.AsyncClient(timeout=timeout) as client:
+        async with client.stream("GET", url) as response:
+            print(f"SSE connected, status: {response.status_code}")
+
+            if response.status_code != 200:
+                print(f"Error: SSE endpoint returned {response.status_code}")
+                return events
+
+            async for line in response.aiter_lines():
+                if not line:
+                    continue
+
+                print(f"  Raw SSE: {line[:100]}...")  # Print first 100 chars
+
+                if line.startswith("data: "):
+                    try:
+                        data = json.loads(line[6:])
+                        print(f"  → {data['status']:12} {data.get('progress', 0):6.1f}% {data.get('filename', '')}")
+                        events.append(data)
+
+                        # Stop if complete or error
+                        if data.get("status") in ("complete", "error"):
+                            print(f"  Download {data['status']}!")
+                            break
+
+                    except json.JSONDecodeError as e:
+                        print(f"  Error parsing JSON: {e}")
+                        print(f"  Line was: {line}")
+
+                elif line.startswith(": heartbeat"):
+                    print("  ♥ heartbeat")
+
+    return events
+
+
+async def trigger_download(model_name: str):
+    """Trigger a model download via the API."""
+    url = "http://localhost:8000/models/download"
+
+    print(f"\nTriggering download for: {model_name}")
+
+    async with httpx.AsyncClient(timeout=300) as client:
+        response = await client.post(url, json={"model_name": model_name})
+        print(f"Response: {response.status_code} - {response.json()}")
+        return response.status_code == 200
+
+
+async def check_server():
+    """Check if the server is running."""
+    try:
+        async with httpx.AsyncClient(timeout=5) as client:
+            response = await client.get("http://localhost:8000/health")
+            return response.status_code == 200
+    except Exception as e:
+        print(f"Server not running: {e}")
+        return False
+
+
+async def main():
+    print("=" * 60)
+    print("Real Model Download Progress Test")
+    print("=" * 60)
+
+    # Check if server is running
+    print("\nChecking if server is running...")
+    if not await check_server():
+        print("✗ Server is not running on http://localhost:8000")
+        print("\nPlease start the server first:")
+        print("  cd backend && python main.py")
+        return False
+
+    print("✓ Server is running")
+
+    # Choose a small model for testing
+    model_name = "whisper-base"  # ~150MB, faster to download
+    print(f"\nUsing model: {model_name}")
+
+    # Option to delete model first if it exists
+    print("\nDo you want to delete the model first to force a fresh download? (y/n)")
+    # For automated testing, skip deletion prompt
+    # delete_first = input().strip().lower() == 'y'
+    delete_first = False
+
+    if delete_first:
+        print(f"Deleting {model_name}...")
+        async with httpx.AsyncClient(timeout=30) as client:
+            response = await client.delete(f"http://localhost:8000/models/{model_name}")
+            print(f"Delete response: {response.status_code}")
+
+    print("\n" + "=" * 60)
+    print("Starting Test")
+    print("=" * 60)
+
+    # Start monitoring SSE stream BEFORE triggering download
+    async def run_test():
+        # Start SSE monitor in background
+        monitor_task = asyncio.create_task(monitor_sse_stream(model_name))
+
+        # Wait a bit to ensure SSE is connected
+        await asyncio.sleep(1)
+
+        # Trigger download
+        success = await trigger_download(model_name)
+
+        if not success:
+            print("✗ Failed to trigger download")
+            monitor_task.cancel()
+            return False
+
+        # Wait for SSE monitor to complete
+        events = await monitor_task
+
+        return events
+
+    events = await run_test()
+
+    # Results
+    print("\n" + "=" * 60)
+    print("Test Results")
+    print("=" * 60)
+
+    if not events:
+        print("✗ FAILED - No SSE events received!")
+        print("\nPossible causes:")
+        print("  1. SSE endpoint not working")
+        print("  2. Progress updates not being sent")
+        print("  3. Model already downloaded (no progress to report)")
+        print("\nTry deleting the model first to force a fresh download:")
+        print(f"  curl -X DELETE http://localhost:8000/models/{model_name}")
+        return False
+
+    print(f"✓ Received {len(events)} SSE events")
+    print(f"\nFirst event: {events[0]}")
+    print(f"Last event: {events[-1]}")
+
+    # Check if we got meaningful progress
+    has_progress = any(e.get('progress', 0) > 0 for e in events)
+    has_complete = any(e.get('status') == 'complete' for e in events)
+
+    if has_progress:
+        print("✓ Progress updates received")
+    else:
+        print("✗ No progress updates (might be already downloaded)")
+
+    if has_complete:
+        print("✓ Download completed successfully")
+    else:
+        print("✗ Download did not complete")
+
+    success = has_progress and has_complete
+
+    if success:
+        print("\n✓ TEST PASSED - Progress tracking works!")
+    else:
+        print("\n⊘ TEST INCONCLUSIVE - Try with a fresh download")
+
+    return success
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

backend/utils/hf_progress.py

@@ -11,8 +11,9 @@ import sys
 class HFProgressTracker:
     """Tracks HuggingFace Hub download progress by intercepting tqdm."""
     
-    def __init__(self, progress_callback: Optional[Callable] = None):
+    def __init__(self, progress_callback: Optional[Callable] = None, filter_non_downloads: bool = False):
         self.progress_callback = progress_callback
+        self.filter_non_downloads = filter_non_downloads  # Only filter if True
         self._original_tqdm_class = None
         self._lock = threading.Lock()
         self._total_downloaded = 0
@@ -21,6 +22,7 @@ class HFProgressTracker:
         self._file_downloaded = {}  # Track downloaded bytes per file
         self._current_filename = ""
         self._active_tqdms = {}  # Track active tqdm instances
+        self._hf_tqdm_original_update = None  # For monkey-patching hf's tqdm
     
     def _create_tracked_tqdm_class(self):
         """Create a tqdm subclass that tracks progress."""
@@ -31,7 +33,6 @@ class HFProgressTracker:
             """A tqdm subclass that reports progress to our tracker."""
 
             def __init__(self, *args, **kwargs):
-                print(f"[DEBUG TrackedTqdm] __init__ called with desc: {kwargs.get('desc', '')}")
                 # Extract filename from desc before passing to parent
                 desc = kwargs.get("desc", "")
                 if not desc and args:
@@ -80,7 +81,6 @@ class HFProgressTracker:
                     }
             
             def update(self, n=1):
-                print(f"[DEBUG TrackedTqdm] update called with n={n}")
                 result = super().update(n)
 
                 # Report progress
@@ -91,6 +91,16 @@ class HFProgressTracker:
                         total = getattr(self, "total", 0)
                         
                         if total and total > 0:
+                            # Always filter out non-byte progress bars (e.g., "Fetching 12 files")
+                            # These cause crazy percentages because they're counting files, not bytes
+                            if self._is_non_byte_progress(filename):
+                                return result
+                            
+                            # When model is cached, also filter out generation-related progress
+                            if tracker.filter_non_downloads:
+                                if not self._is_download_progress(filename):
+                                    return result
+                            
                             # Update per-file tracking
                             tracker._file_sizes[filename] = total
                             tracker._file_downloaded[filename] = current
@@ -99,6 +109,13 @@ class HFProgressTracker:
                             tracker._total_size = sum(tracker._file_sizes.values())
                             tracker._total_downloaded = sum(tracker._file_downloaded.values())
                             
+                            # Only report progress once we have a meaningful total (at least 1MB)
+                            # This avoids the "100% at 0MB" issue when small config
+                            # files are counted before the real model files
+                            MIN_TOTAL_BYTES = 1_000_000  # 1MB
+                            if tracker._total_size < MIN_TOTAL_BYTES:
+                                return result
+                            
                             # Call progress callback
                             if tracker.progress_callback:
                                 tracker.progress_callback(
@@ -109,6 +126,50 @@ class HFProgressTracker:
                 
                 return result
             
+            def _is_non_byte_progress(self, filename: str) -> bool:
+                """Check if this progress bar should be SKIPPED (returns True to skip).
+                
+                We want to track byte-based progress bars. This method identifies
+                progress bars that count files/items instead of bytes, which would
+                cause crazy percentages if mixed with our byte counting.
+                
+                Returns:
+                    True = SKIP this bar (it's not byte-based)
+                    False = TRACK this bar (it counts bytes)
+                """
+                if not filename:
+                    return False
+                
+                filename_lower = filename.lower()
+                
+                # Skip "Fetching X files" - it counts files (total=12), not bytes
+                # Don't skip "Downloading (incomplete total...)" - that IS byte-based
+                skip_patterns = [
+                    'fetching',  # "Fetching 12 files" has total=12 files, not bytes
+                ]
+                return any(pattern in filename_lower for pattern in skip_patterns)
+            
+            def _is_download_progress(self, filename: str) -> bool:
+                """Check if this is a real file download progress bar vs internal processing."""
+                if not filename or filename == "unknown":
+                    return False
+                
+                # Real downloads have file extensions
+                download_extensions = [
+                    '.safetensors', '.bin', '.pt', '.pth',  # Model weights
+                    '.json', '.txt', '.py',  # Config files
+                    '.msgpack', '.h5',  # Other formats
+                ]
+                
+                filename_lower = filename.lower()
+                has_extension = any(filename_lower.endswith(ext) for ext in download_extensions)
+                
+                # Skip generation-related progress indicators
+                skip_patterns = ['segment', 'processing', 'generating', 'loading']
+                has_skip_pattern = any(pattern in filename_lower for pattern in skip_patterns)
+                
+                return has_extension and not has_skip_pattern
+            
             def close(self):
                 with tracker._lock:
                     if id(self) in tracker._active_tqdms:
@@ -120,13 +181,11 @@ class HFProgressTracker:
     @contextmanager
     def patch_download(self):
         """Context manager to patch tqdm for progress tracking."""
-        print("[DEBUG HFProgressTracker] patch_download called")
         try:
             import tqdm as tqdm_module
 
             # Store original tqdm class
             self._original_tqdm_class = tqdm_module.tqdm
-            print(f"[DEBUG HFProgressTracker] Original tqdm class: {self._original_tqdm_class}")
             
             # Reset totals
             with self._lock:
@@ -139,39 +198,89 @@ class HFProgressTracker:
             
             # Create our tracked tqdm class
             tracked_tqdm = self._create_tracked_tqdm_class()
-            print(f"[DEBUG HFProgressTracker] Created TrackedTqdm class: {tracked_tqdm}")
 
             # Patch tqdm.tqdm
             tqdm_module.tqdm = tracked_tqdm
-            print(f"[DEBUG HFProgressTracker] Patched tqdm.tqdm")
 
             # Also patch tqdm.auto.tqdm if it exists (used by huggingface_hub)
             self._original_tqdm_auto = None
             if hasattr(tqdm_module, "auto") and hasattr(tqdm_module.auto, "tqdm"):
                 self._original_tqdm_auto = tqdm_module.auto.tqdm
                 tqdm_module.auto.tqdm = tracked_tqdm
-                print(f"[DEBUG HFProgressTracker] Patched tqdm.auto.tqdm")
             
             # Patch in sys.modules to catch already-imported references
+            # huggingface_hub uses: from tqdm.auto import tqdm as base_tqdm
+            # So we need to patch both 'tqdm' and 'base_tqdm' attributes
             self._patched_modules = {}
+            tqdm_attr_names = ['tqdm', 'base_tqdm', 'old_tqdm']  # Various names used
+            
             patched_count = 0
             for module_name in list(sys.modules.keys()):
                 if "huggingface" in module_name or module_name.startswith("tqdm"):
                     try:
                         module = sys.modules[module_name]
-                        if hasattr(module, "tqdm"):
+                        for attr_name in tqdm_attr_names:
-                            attr = getattr(module, "tqdm")
+                            if hasattr(module, attr_name):
-                            # Only patch if it's the original tqdm class (not already patched)
+                                attr = getattr(module, attr_name)
-                            if attr is self._original_tqdm_class or (
+                                # Only patch if it's a tqdm class (not already patched)
-                                hasattr(attr, "__name__") and attr.__name__ == "tqdm"
+                                is_tqdm_class = (
-                            ):
+                                    attr is self._original_tqdm_class or 
-                                self._patched_modules[module_name] = attr
+                                    (self._original_tqdm_auto and attr is self._original_tqdm_auto) or
-                                setattr(module, "tqdm", tracked_tqdm)
+                                    (hasattr(attr, "__name__") and attr.__name__ == "tqdm" and 
+                                     hasattr(attr, "update"))  # tqdm classes have update method
+                                )
+                                if is_tqdm_class:
+                                    key = f"{module_name}.{attr_name}"
+                                    self._patched_modules[key] = (module, attr_name, attr)
+                                    setattr(module, attr_name, tracked_tqdm)
                                     patched_count += 1
-                                print(f"[DEBUG HFProgressTracker] Patched {module_name}.tqdm")
                     except (AttributeError, TypeError):
                         pass
-            print(f"[DEBUG HFProgressTracker] Patched {patched_count} modules in sys.modules")
+            
+            # ALSO monkey-patch the update method on huggingface_hub's tqdm class
+            # This is needed because the class was already defined at import time
+            self._hf_tqdm_original_update = None
+            try:
+                from huggingface_hub.utils import tqdm as hf_tqdm_module
+                if hasattr(hf_tqdm_module, 'tqdm'):
+                    hf_tqdm_class = hf_tqdm_module.tqdm
+                    self._hf_tqdm_original_update = hf_tqdm_class.update
+                    
+                    # Create a wrapper that calls our tracking
+                    tracker = self  # Reference to HFProgressTracker instance
+                    def patched_update(tqdm_self, n=1):
+                        result = tracker._hf_tqdm_original_update(tqdm_self, n)
+                        
+                        # Track this progress
+                        with tracker._lock:
+                            desc = getattr(tqdm_self, 'desc', '') or ''
+                            current = getattr(tqdm_self, 'n', 0)
+                            total = getattr(tqdm_self, 'total', 0) or 0
+                            
+                            # Skip non-byte progress bars
+                            if 'fetching' in desc.lower():
+                                return result
+                            
+                            # Skip until we have a meaningful total (at least 1MB)
+                            # This avoids the "100% at 0MB" issue when small config
+                            # files are counted before the real model files
+                            MIN_TOTAL_BYTES = 1_000_000  # 1MB
+                            if total >= MIN_TOTAL_BYTES:
+                                tracker._total_downloaded = current
+                                tracker._total_size = total
+                                
+                                if tracker.progress_callback:
+                                    tracker.progress_callback(current, total, desc)
+                        
+                        return result
+                    
+                    hf_tqdm_class.update = patched_update
+                    patched_count += 1
+                    print(f"[HFProgressTracker] Monkey-patched huggingface_hub.utils.tqdm.tqdm.update")
+            except (ImportError, AttributeError) as e:
+                print(f"[HFProgressTracker] Could not monkey-patch hf_tqdm: {e}")
+            
+            print(f"[HFProgressTracker] Patched {patched_count} tqdm references")
             
             yield
             
@@ -189,15 +298,24 @@ class HFProgressTracker:
                         tqdm_module.auto.tqdm = self._original_tqdm_auto
                     
                     # Restore patched modules
-                    for module_name, original in self._patched_modules.items():
+                    for key, (module, attr_name, original) in self._patched_modules.items():
                         try:
-                            module = sys.modules.get(module_name)
                             if module and original:
-                                setattr(module, "tqdm", original)
+                                setattr(module, attr_name, original)
                         except (AttributeError, TypeError):
                             pass
                     self._patched_modules = {}
                     
+                    # Restore hf_tqdm's original update method
+                    if self._hf_tqdm_original_update:
+                        try:
+                            from huggingface_hub.utils import tqdm as hf_tqdm_module
+                            if hasattr(hf_tqdm_module, 'tqdm'):
+                                hf_tqdm_module.tqdm.update = self._hf_tqdm_original_update
+                        except (ImportError, AttributeError):
+                            pass
+                        self._hf_tqdm_original_update = None
+                    
                 except (ImportError, AttributeError):
                     pass
 
@@ -205,8 +323,12 @@ class HFProgressTracker:
 def create_hf_progress_callback(model_name: str, progress_manager):
     """Create a progress callback for HuggingFace downloads."""
     def callback(downloaded: int, total: int, filename: str = ""):
-        """Progress callback."""
+        """Progress callback.
-        if total > 0:
+        
+        Note: We send updates even when total=0 (unknown) to provide feedback
+        during the "incomplete total" phase of huggingface_hub downloads.
+        The frontend handles total=0 gracefully.
+        """
         progress_manager.update_progress(
             model_name=model_name,
             current=downloaded,

backend/utils/progress.py

@@ -16,11 +16,17 @@ class ProgressManager:
     Thread-safe: can be called from background threads (e.g., via asyncio.to_thread).
     """
     
+    # Throttle settings to prevent overwhelming SSE clients
+    THROTTLE_INTERVAL_SECONDS = 0.5  # Minimum time between updates
+    THROTTLE_PROGRESS_DELTA = 1.0    # Minimum progress change (%) to force update
+    
     def __init__(self):
         self._progress: Dict[str, Dict] = {}
         self._listeners: Dict[str, list] = {}
         self._lock = threading.Lock()  # Thread-safe lock for progress dict
         self._main_loop: Optional[asyncio.AbstractEventLoop] = None
+        self._last_notify_time: Dict[str, float] = {}  # Last notification time per model
+        self._last_notify_progress: Dict[str, float] = {}  # Last notified progress per model
     
     def _set_main_loop(self, loop: asyncio.AbstractEventLoop):
         """Set the main event loop for thread-safe operations."""
@@ -68,6 +74,10 @@ class ProgressManager:
 
         Thread-safe: can be called from background threads.
         
+        Progress updates are throttled to prevent overwhelming SSE clients.
+        Updates are sent at most every THROTTLE_INTERVAL_SECONDS, or when
+        progress changes by at least THROTTLE_PROGRESS_DELTA percent.
+
         Args:
             model_name: Name of the model (e.g., "qwen-tts-1.7B", "whisper-base")
             current: Current bytes downloaded
@@ -76,9 +86,17 @@ class ProgressManager:
             status: Status string (downloading, extracting, complete, error)
         """
         import logging
+        import time
         logger = logging.getLogger(__name__)
 
-        progress_pct = (current / total * 100) if total > 0 else 0
+        # Calculate progress percentage, clamped to 0-100 range
+        # This prevents crazy percentages from edge cases like:
+        # - current > total temporarily during aggregation
+        # - mixing file-count progress with byte-count progress
+        if total > 0:
+            progress_pct = min(100.0, max(0.0, (current / total * 100)))
+        else:
+            progress_pct = 0
 
         progress_data = {
             "model_name": model_name,
@@ -90,25 +108,38 @@ class ProgressManager:
             "timestamp": datetime.now().isoformat(),
         }
 
-        print(f"[DEBUG] update_progress called: {model_name}, {progress_pct:.1f}%")
+        # Thread-safe update of progress dict (always update internal state)
-
-        # Thread-safe update of progress dict
         with self._lock:
             self._progress[model_name] = progress_data
 
+        # Check if we should notify listeners (throttling)
+        current_time = time.time()
+        last_time = self._last_notify_time.get(model_name, 0)
+        last_progress = self._last_notify_progress.get(model_name, -100)
+        
+        time_delta = current_time - last_time
+        progress_delta = abs(progress_pct - last_progress)
+        
+        # Always notify for complete/error status, or if throttle conditions are met
+        should_notify = (
+            status in ("complete", "error") or
+            time_delta >= self.THROTTLE_INTERVAL_SECONDS or
+            progress_delta >= self.THROTTLE_PROGRESS_DELTA
+        )
+        
+        if not should_notify:
+            return  # Skip this update (throttled)
+        
+        # Update throttle tracking
+        self._last_notify_time[model_name] = current_time
+        self._last_notify_progress[model_name] = progress_pct
+
         # Notify all listeners (thread-safe)
         listener_count = len(self._listeners.get(model_name, []))
-        print(f"[DEBUG] Listener count for {model_name}: {listener_count}")
-        print(f"[DEBUG] All listeners: {list(self._listeners.keys())}")
-        print(f"[DEBUG] Main loop set: {self._main_loop is not None}")
-        if self._main_loop:
-            print(f"[DEBUG] Main loop running: {self._main_loop.is_running()}")
 
         if listener_count > 0:
             logger.debug(f"Notifying {listener_count} listeners for {model_name}: {progress_pct:.1f}% ({filename})")
-            print(f"[DEBUG] About to notify listeners...")
             self._notify_listeners_threadsafe(model_name, progress_data)
-            print(f"[DEBUG] Notified listeners")
         else:
             logger.debug(f"No listeners for {model_name}, progress update stored: {progress_pct:.1f}%")
     

backend/voicebox-server-cuda.spec

@@ -0,0 +1,48 @@
+# -*- mode: python ; coding: utf-8 -*-
+from PyInstaller.utils.hooks import collect_data_files
+from PyInstaller.utils.hooks import collect_submodules
+from PyInstaller.utils.hooks import copy_metadata
+
+datas = []
+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', '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', 'pkg_resources.extern']
+datas += collect_data_files('qwen_tts')
+datas += copy_metadata('qwen-tts')
+hiddenimports += collect_submodules('qwen_tts')
+hiddenimports += collect_submodules('jaraco')
+
+
+a = Analysis(
+    ['server.py'],
+    pathex=[],
+    binaries=[],
+    datas=datas,
+    hiddenimports=hiddenimports,
+    hookspath=[],
+    hooksconfig={},
+    runtime_hooks=[],
+    excludes=[],
+    noarchive=False,
+    optimize=0,
+)
+pyz = PYZ(a.pure)
+
+exe = EXE(
+    pyz,
+    a.scripts,
+    a.binaries,
+    a.datas,
+    [],
+    name='voicebox-server-cuda',
+    debug=False,
+    bootloader_ignore_signals=False,
+    strip=False,
+    upx=True,
+    upx_exclude=[],
+    runtime_tmpdir=None,
+    console=True,
+    disable_windowed_traceback=False,
+    argv_emulation=False,
+    target_arch=None,
+    codesign_identity=None,
+    entitlements_file=None,
+)

backend/voicebox-server.spec

@@ -4,15 +4,11 @@ from PyInstaller.utils.hooks import collect_submodules
 from PyInstaller.utils.hooks import copy_metadata
 
 datas = []
-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', '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', '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.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', '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', 'pkg_resources.extern']
 datas += collect_data_files('qwen_tts')
-datas += collect_data_files('mlx')
-datas += collect_data_files('mlx_audio')
 datas += copy_metadata('qwen-tts')
 hiddenimports += collect_submodules('qwen_tts')
 hiddenimports += collect_submodules('jaraco')
-hiddenimports += collect_submodules('mlx')
-hiddenimports += collect_submodules('mlx_audio')
 
 
 a = Analysis(

bun.lock

@@ -13,7 +13,7 @@
     },
     "app": {
       "name": "@voicebox/app",
-      "version": "0.1.9",
+      "version": "0.1.11",
       "dependencies": {
         "@dnd-kit/core": "^6.3.1",
         "@dnd-kit/sortable": "^10.0.0",
@@ -68,7 +68,7 @@
     },
     "landing": {
       "name": "@voicebox/landing",
-      "version": "0.1.9",
+      "version": "0.1.11",
       "dependencies": {
         "@radix-ui/react-separator": "^1.1.8",
         "@radix-ui/react-slot": "^1.2.4",
@@ -93,10 +93,14 @@
     },
     "tauri": {
       "name": "@voicebox/tauri",
-      "version": "0.1.9",
+      "version": "0.1.11",
       "dependencies": {
         "@tauri-apps/api": "^2.0.0",
+        "@tauri-apps/plugin-dialog": "^2.0.0",
+        "@tauri-apps/plugin-fs": "^2.0.0",
+        "@tauri-apps/plugin-process": "^2.0.0",
         "@tauri-apps/plugin-shell": "^2.0.0",
+        "@tauri-apps/plugin-updater": "^2.0.0",
       },
       "devDependencies": {
         "@tailwindcss/vite": "^4.1.18",
@@ -112,7 +116,7 @@
     },
     "web": {
       "name": "@voicebox/web",
-      "version": "0.1.9",
+      "version": "0.1.11",
       "dependencies": {
         "@tanstack/react-query": "^5.0.0",
         "react": "^18.3.0",

docs/cuda-distribution-problem.md

@@ -0,0 +1,620 @@
+# CUDA Distribution Problem - Complete Analysis
+
+## Table of Contents
+1. [Problem Overview](#problem-overview)
+2. [Root Cause](#root-cause)
+3. [Attempted Solutions](#attempted-solutions)
+4. [Current Status](#current-status)
+5. [Available Options](#available-options)
+6. [Technical Details](#technical-details)
+7. [Cost Analysis](#cost-analysis)
+8. [Recommendations](#recommendations)
+
+---
+
+## Problem Overview
+
+### Timeline of Issues
+
+**Original Problem (v0.1.0 - v0.1.11)**
+- Single server binary with CUDA support
+- Size: ~2.9GB
+- Issue: MSI installer build fails in GitHub Actions CI
+- Error: WiX Toolset cannot handle 3GB files efficiently
+
+**First Solution: Dual Binary System (v0.1.12)**
+- Split into CPU (295MB) and CUDA (2.37GB) binaries
+- CPU ships with installer
+- CUDA as optional download
+- Issue: GitHub Release assets have 2GB limit
+
+**Current Problem (Discovered during implementation)**
+- GitHub Release Asset Limit: **2GB hard maximum**
+- CUDA binary: **2.37GB** (370MB over limit)
+- Cannot upload to GitHub Releases
+
+---
+
+## Root Cause
+
+### Why Is The CUDA Binary So Large?
+
+The size difference between CPU and CUDA builds:
+
+| Component | CPU Build | CUDA Build | Difference |
+|-----------|-----------|------------|------------|
+| PyTorch Core | ~150MB | ~150MB | - |
+| CPU Libraries (MKL/OpenBLAS) | ~100MB | - | -100MB |
+| CUDA Runtime | - | ~500MB | +500MB |
+| cuBLAS | - | ~350MB | +350MB |
+| cuDNN | - | ~1.2GB | +1.2GB |
+| NVRTC (CUDA Compiler) | - | ~90MB | +90MB |
+| Other CUDA libs | - | ~100MB | +100MB |
+| **Total** | **~295MB** | **~2.37GB** | **+2.07GB** |
+
+### CUDA Dependencies Breakdown
+
+```
+torch/lib/ (CUDA build):
+├── cudart64_12.dll          (~0.5 MB)   - CUDA Runtime
+├── cublas64_12.dll          (~100 MB)   - Basic Linear Algebra
+├── cublasLt64_12.dll        (~200 MB)   - Linear Algebra (optimized)
+├── cudnn64_9.dll            (~800 MB)   - Deep Neural Networks
+├── cudnn_*_infer64_9.dll    (~400 MB)   - DNN Inference ops
+├── nvrtc64_*.dll            (~50 MB)    - Runtime Compiler
+├── nvrtc-builtins64_*.dll   (~40 MB)    - Compiler builtins
+├── torch_cuda.dll           (~200 MB)   - PyTorch CUDA bridge
+└── c10_cuda.dll             (~20 MB)    - Core CUDA utilities
+```
+
+**Why These Are Required:**
+- cuDNN is essential for neural network operations
+- cuBLAS handles all matrix operations (core of ML)
+- Cannot split or remove without breaking functionality
+
+---
+
+## Attempted Solutions
+
+### Solution 1: Dual Binary System ✅ (Partially Successful)
+
+**Goal**: Split CPU and CUDA into separate downloads
+
+**Implementation**:
+```bash
+# Build CPU-only (295MB)
+pip install torch --index-url https://download.pytorch.org/whl/cpu
+python build_binary.py cpu
+
+# Build CUDA (2.37GB)
+pip install torch --index-url https://download.pytorch.org/whl/cu121
+python build_binary.py cuda
+```
+
+**Results**:
+- ✅ CPU binary: 295MB (fits in installer)
+- ✅ CI builds successfully
+- ✅ Installer size reduced from 3GB to ~500MB
+- ❌ CUDA binary still too large for GitHub
+
+**See**: `docs/dual-server-binaries.md`
+
+### Solution 2: Compression Testing ❌ (Failed)
+
+**Goal**: Compress CUDA binary to fit under 2GB
+
+**Method**: 7z with maximum compression settings
+```bash
+7z a -t7z -m0=lzma2 -mx=9 -mfb=64 -md=32m -ms=on \
+  voicebox-server-cuda.7z voicebox-server-cuda.exe
+```
+
+**Results**:
+```
+Original:     2.37 GB (2,545,086,396 bytes)
+Compressed:   2.35 GB (2,519,381,264 bytes)
+Compression:  1.0% (only 24.5MB saved)
+GitHub Limit: 2.00 GB (2,147,483,648 bytes)
+Over by:      354.67 MB
+
+Status: FAILED - Still exceeds limit by 354MB
+```
+
+**Why Compression Failed**:
+- CUDA binaries are already optimized machine code
+- No redundant data to compress
+- Neural network kernels are highly compact
+- Libraries are already stripped of debug symbols
+
+**Conclusion**: Compression is not viable
+
+---
+
+## Current Status
+
+### What Works
+- ✅ CPU binary builds successfully (295MB)
+- ✅ CUDA binary builds successfully (2.37GB)
+- ✅ Build scripts for both variants
+- ✅ CI workflow updated for dual binaries
+- ✅ Installer can be created with CPU binary
+
+### What Doesn't Work
+- ❌ Cannot upload CUDA binary to GitHub Releases (exceeds 2GB limit)
+- ❌ Compression doesn't reduce size enough
+- ❌ No automated distribution path for CUDA binary
+
+### Branch Status
+- Branch: `feat/dual-server-binaries`
+- Commits: Implementation complete
+- Testing: Local builds successful
+- Blocker: CUDA distribution path
+
+---
+
+## Available Options
+
+### Option 1: AWS S3 Hosting (Recommended)
+
+**Description**: Host CUDA binary in Amazon S3 bucket
+
+**Pros**:
+- ✅ No file size limits (can handle multi-GB files)
+- ✅ Fast global CDN (CloudFront)
+- ✅ Reliable (99.99% uptime)
+- ✅ Pay only for usage
+- ✅ Easy CI integration
+- ✅ Version control (keep multiple releases)
+
+**Cons**:
+- ❌ Requires AWS account
+- ❌ Monthly costs (~$1-5/month)
+- ❌ Additional infrastructure to manage
+
+**Cost Estimate**:
+```
+Storage: 2.37 GB × $0.023/GB = $0.05/month
+Transfer: 100 downloads × 2.37GB × $0.09/GB = $21.33/month
+Total: ~$21-25/month for 100 downloads
+      ~$2-5/month for 10-20 downloads
+```
+
+**Implementation**:
+```yaml
+# .github/workflows/release.yml
+- name: Upload CUDA to S3
+  env:
+    AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
+    AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
+  run: |
+    aws s3 cp backend/cuda-release/voicebox-server-cuda-*.exe \
+      s3://voicebox-releases/cuda/${{ github.ref_name }}/ \
+      --acl public-read
+
+    # Generate download URL
+    echo "CUDA_URL=https://voicebox-releases.s3.amazonaws.com/cuda/${{ github.ref_name }}/voicebox-server-cuda-x86_64-pc-windows-msvc.exe" >> release_notes.txt
+```
+
+**User Experience**:
+1. Install app normally (500MB installer)
+2. App detects NVIDIA GPU
+3. Shows: "Download CUDA support? (2.4GB)"
+4. Downloads from S3: `https://voicebox-releases.s3.amazonaws.com/cuda/v0.1.12/voicebox-server-cuda.exe`
+5. Saves to `%APPDATA%/voicebox/binaries/`
+6. App restarts with CUDA server
+
+---
+
+### Option 2: Azure Blob Storage
+
+**Description**: Microsoft Azure alternative to S3
+
+**Pros**:
+- ✅ Similar to S3 (no size limits, CDN, reliable)
+- ✅ Good if already using Azure
+- ✅ Competitive pricing
+- ✅ Global CDN with Azure CDN
+
+**Cons**:
+- ❌ Requires Azure account
+- ❌ Similar monthly costs
+- ❌ Less common in open source projects
+
+**Cost Estimate**:
+```
+Storage: $0.018/GB = $0.04/month
+Transfer: ~$20-25/month for 100 downloads
+```
+
+**Implementation**:
+```yaml
+- name: Upload to Azure Blob
+  env:
+    AZURE_STORAGE_CONNECTION_STRING: ${{ secrets.AZURE_STORAGE }}
+  run: |
+    az storage blob upload \
+      --account-name voiceboxreleases \
+      --container-name cuda-binaries \
+      --name v${{ github.ref_name }}/voicebox-server-cuda.exe \
+      --file backend/cuda-release/voicebox-server-cuda-*.exe \
+      --tier Hot
+```
+
+---
+
+### Option 3: Cloudflare R2
+
+**Description**: Cloudflare's S3-compatible object storage
+
+**Pros**:
+- ✅ S3-compatible API
+- ✅ **FREE egress (no bandwidth charges!)**
+- ✅ Cheaper than S3/Azure
+- ✅ Cloudflare CDN included
+- ✅ Good for open source projects
+
+**Cons**:
+- ❌ Requires Cloudflare account
+- ❌ Newer service (less mature than S3)
+
+**Cost Estimate**:
+```
+Storage: $0.015/GB = $0.04/month
+Egress: $0.00 (FREE!)
+Class A ops: Negligible
+Total: ~$0.04/month (essentially free!)
+```
+
+**Why This Is Attractive**:
+- Zero bandwidth costs (huge savings)
+- Perfect for open source distribution
+- S3-compatible (easy migration if needed)
+
+**Implementation**:
+Same as S3 (R2 is S3-compatible):
+```yaml
+- name: Upload to R2
+  env:
+    AWS_ACCESS_KEY_ID: ${{ secrets.R2_ACCESS_KEY_ID }}
+    AWS_SECRET_ACCESS_KEY: ${{ secrets.R2_SECRET_ACCESS_KEY }}
+    AWS_ENDPOINT_URL: https://<account-id>.r2.cloudflarestorage.com
+  run: |
+    aws s3 cp backend/cuda-release/voicebox-server-cuda-*.exe \
+      s3://voicebox-releases/cuda/${{ github.ref_name }}/ \
+      --endpoint-url=$AWS_ENDPOINT_URL
+```
+
+---
+
+### Option 4: GitHub Packages (Container Registry)
+
+**Description**: Package CUDA binary as OCI/Docker artifact
+
+**Pros**:
+- ✅ Stays in GitHub ecosystem
+- ✅ No additional accounts needed
+- ✅ Free for public repos
+
+**Cons**:
+- ❌ Complex for desktop app distribution
+- ❌ Users need to extract from container
+- ❌ Awkward UX (not designed for binary distribution)
+- ❌ Requires Docker understanding
+
+**Not Recommended**: Containers aren't designed for desktop app binaries
+
+---
+
+### Option 5: Self-Hosted Server
+
+**Description**: Host on your own VPS/server
+
+**Pros**:
+- ✅ Full control
+- ✅ No cloud provider dependency
+- ✅ Predictable costs
+
+**Cons**:
+- ❌ Requires server maintenance
+- ❌ Bandwidth costs can be high
+- ❌ Uptime responsibility
+- ❌ Scaling challenges
+
+**Cost Estimate**:
+```
+VPS: $5-20/month (DigitalOcean, Linode)
+Bandwidth: $0.01-0.02/GB
+Total: $10-50/month depending on traffic
+```
+
+---
+
+### Option 6: Manual Distribution
+
+**Description**: Don't automate - provide manual download instructions
+
+**Pros**:
+- ✅ Zero cost
+- ✅ Zero infrastructure
+- ✅ Simple
+
+**Cons**:
+- ❌ Poor user experience
+- ❌ Manual upload to file host each release
+- ❌ Users must manually download and install
+- ❌ No automatic updates for CUDA binary
+- ❌ Increases support burden
+
+**Implementation**:
+```
+Release notes:
+"Windows users with NVIDIA GPUs can download CUDA support:
+1. Download voicebox-server-cuda.exe from [Google Drive/Mega/etc]
+2. Place in C:\Users\<YourName>\AppData\Roaming\voicebox\binaries\
+3. Restart the app"
+```
+
+**Not Recommended**: Creates friction, support issues
+
+---
+
+### Option 7: Split CUDA Binary
+
+**Description**: Break CUDA binary into multiple <2GB chunks
+
+**Technical Approach**:
+```python
+# Split binary
+split -b 2000M voicebox-server-cuda.exe cuda_part_
+
+# Upload parts to GitHub (each <2GB)
+cuda_part_aa (2.0 GB)
+cuda_part_ab (0.37 GB)
+
+# App downloads and reassembles
+cat cuda_part_* > voicebox-server-cuda.exe
+```
+
+**Pros**:
+- ✅ Stays on GitHub
+- ✅ No external hosting
+
+**Cons**:
+- ❌ Complex download logic (multiple files)
+- ❌ Integrity checking required
+- ❌ More points of failure
+- ❌ Users must wait for multiple downloads
+- ❌ Still hacky solution
+
+**Complexity**: Medium-High
+
+---
+
+## Technical Details
+
+### Current Build Output
+
+```
+backend/dist/
+├── voicebox-server.exe          295 MB  (CPU-only)
+└── voicebox-server-cuda.exe    2.37 GB  (CUDA)
+
+# After compression test:
+backend/dist/
+└── voicebox-server-cuda.7z     2.35 GB  (not viable)
+```
+
+### CI Workflow Changes Required
+
+For external hosting (S3/R2/Azure):
+
+```yaml
+# Current workflow (fails)
+- name: Upload CUDA server binary (Windows only)
+  if: matrix.platform == 'windows-latest'
+  uses: softprops/action-gh-release@v1
+  with:
+    files: backend/cuda-release/voicebox-server-cuda-*.exe  # ❌ Fails: >2GB
+    draft: true
+
+# New workflow (S3 example)
+- name: Upload CUDA to S3 (Windows only)
+  if: matrix.platform == 'windows-latest'
+  env:
+    AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
+    AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
+  run: |
+    aws s3 cp backend/cuda-release/voicebox-server-cuda-*.exe \
+      s3://voicebox-releases/cuda/${{ github.ref_name }}/ \
+      --acl public-read
+
+    # Generate release notes with download URL
+    cat >> release_notes.md <<EOF
+
+    ### GPU Acceleration (Windows)
+    Download CUDA support for NVIDIA GPUs:
+    [voicebox-server-cuda.exe](https://voicebox-releases.s3.amazonaws.com/cuda/${{ github.ref_name }}/voicebox-server-cuda-x86_64-pc-windows-msvc.exe)
+    Size: 2.37 GB
+    EOF
+```
+
+### App Changes Required
+
+**Frontend (Tauri)**: Download manager
+```typescript
+// src/lib/cuda-downloader.ts
+const CUDA_DOWNLOAD_URL =
+  "https://voicebox-releases.s3.amazonaws.com/cuda/v{VERSION}/voicebox-server-cuda.exe";
+
+async function downloadCudaBinary(version: string) {
+  const url = CUDA_DOWNLOAD_URL.replace("{VERSION}", version);
+  const savePath = path.join(app.getPath("userData"), "binaries", "voicebox-server-cuda.exe");
+
+  // Download with progress
+  await downloadFile(url, savePath, (progress) => {
+    // Update UI: "Downloading CUDA support: 45% (1.2GB / 2.4GB)"
+  });
+
+  // Verify checksum
+  const checksum = await calculateChecksum(savePath);
+  if (checksum !== EXPECTED_CHECKSUM) {
+    throw new Error("Download corrupted");
+  }
+}
+```
+
+**Backend**: Already supports both binaries (no changes needed)
+
+---
+
+## Cost Analysis
+
+### Monthly Cost Comparison (100 downloads/month)
+
+| Option | Storage | Bandwidth | Total/Month | Notes |
+|--------|---------|-----------|-------------|-------|
+| **Cloudflare R2** | $0.04 | $0.00 | **$0.04** | Best for open source |
+| AWS S3 | $0.05 | $21.33 | $21.38 | Good reliability |
+| Azure Blob | $0.04 | $20.00 | $20.04 | Azure ecosystem |
+| Self-hosted VPS | $10.00 | $2.37 | $12.37 | Maintenance overhead |
+| Manual | $0.00 | $0.00 | $0.00 | Poor UX |
+
+### Annual Cost Comparison
+
+| Option | Year 1 | Year 2+ | Notes |
+|--------|--------|---------|-------|
+| **Cloudflare R2** | **$0.50** | **$0.50** | Essentially free |
+| AWS S3 | $256 | $256 | Predictable |
+| Self-hosted | $144 | $144 | Time cost |
+
+**Recommendation**: Cloudflare R2 (free egress = huge savings)
+
+---
+
+## Recommendations
+
+### Recommended Solution: Cloudflare R2
+
+**Why**:
+1. **Cost**: Essentially free (~$0.04/month)
+2. **Bandwidth**: Zero egress charges (unlimited downloads)
+3. **CDN**: Cloudflare's global network included
+4. **Compatibility**: S3-compatible API (easy to use)
+5. **Perfect for open source**: No surprise bandwidth bills
+
+### Implementation Priority
+
+**Phase 1: Setup (1-2 hours)**
+1. Create Cloudflare R2 account
+2. Create bucket: `voicebox-releases`
+3. Generate API credentials
+4. Add to GitHub Secrets
+
+**Phase 2: CI Integration (1-2 hours)**
+1. Update `.github/workflows/release.yml`
+2. Add R2 upload step
+3. Generate release notes with download URL
+4. Test with draft release
+
+**Phase 3: App Integration (4-6 hours)**
+1. Add GPU detection on startup
+2. Implement download manager UI
+3. Add progress indicators
+4. Implement checksum verification
+5. Server restart logic
+
+**Phase 4: Documentation (1 hour)**
+1. Update README with GPU instructions
+2. Add troubleshooting guide
+3. Document manual download process
+
+**Total Time**: ~8-12 hours of development
+
+### Alternative: AWS S3 (If Already Using AWS)
+
+If you're already using AWS for other infrastructure, S3 is also a solid choice:
+- More mature than R2
+- Extensive documentation
+- Familiar tooling
+- ~$20/month for moderate usage
+
+---
+
+## Open Questions
+
+1. **Expected Download Volume**: How many CUDA downloads per month?
+   - Affects cost calculations
+   - Determines if R2's free egress is significant
+
+2. **Update Strategy**: How to handle CUDA updates?
+   - Option A: Version in URL path (keep all versions)
+   - Option B: Overwrite latest (save space)
+
+3. **Fallback Strategy**: What if cloud provider is down?
+   - Mirror on multiple providers?
+   - Graceful degradation to CPU?
+
+4. **Telemetry**: Track CUDA download stats?
+   - Helps with cost forecasting
+   - User behavior insights
+
+---
+
+## Next Steps
+
+1. **Research Phase** (You are here)
+   - Evaluate cloud providers
+   - Check terms of service
+   - Test account creation
+
+2. **Decision Phase**
+   - Choose provider (Cloudflare R2 recommended)
+   - Set up account
+   - Configure billing alerts
+
+3. **Implementation Phase**
+   - Update CI workflow
+   - Implement download manager
+   - Test end-to-end flow
+
+4. **Launch Phase**
+   - Deploy to production
+   - Monitor downloads
+   - Gather user feedback
+
+---
+
+## References
+
+- **GitHub Release Limits**: https://docs.github.com/en/repositories/releasing-projects-on-github/about-releases
+- **Cloudflare R2 Pricing**: https://developers.cloudflare.com/r2/pricing/
+- **AWS S3 Pricing**: https://aws.amazon.com/s3/pricing/
+- **Compression Test Results**: `backend/test_cuda_compression.py`
+- **Dual Binary Implementation**: `docs/dual-server-binaries.md`
+
+---
+
+## Appendix: Alternative Approaches Considered
+
+### A. Dynamic CUDA Loading
+**Idea**: Load CUDA DLLs dynamically at runtime
+**Why Not**: PyTorch requires CUDA DLLs at import time, can't lazy-load
+
+### B. CUDA as Separate Package
+**Idea**: Python package with just CUDA libs
+**Why Not**: Still 2GB+, same problem
+
+### C. Model Quantization
+**Idea**: Use smaller quantized models
+**Why Not**: Doesn't reduce CUDA runtime size
+
+### D. Docker Distribution
+**Idea**: Distribute as Docker container
+**Why Not**: Poor fit for desktop app, requires Docker installed
+
+---
+
+**Document Version**: 1.0
+**Last Updated**: 2026-01-31
+**Status**: Research Phase
+**Next Review**: After cloud provider decision

docs/dual-server-binaries.md

@@ -0,0 +1,177 @@
+# Dual Server Binary System
+
+## Overview
+
+Voicebox now uses a dual-binary approach to manage the size difference between CPU-only and CUDA-enabled builds:
+
+- **CPU Binary** (~500MB): Ships with the installer by default
+- **CUDA Binary** (~3GB): Downloaded on-demand for GPU users
+
+## Problem Solved
+
+Previously, bundling PyTorch with CUDA support created a 3GB server binary, which:
+- Made the installer too large (failed CI builds with WiX)
+- Forced all users to download CUDA libraries even without NVIDIA GPUs
+- Created poor user experience
+
+## Solution
+
+### Build Process
+
+**Two separate binaries are built:**
+
+1. **voicebox-server.exe** (CPU)
+   - Built with: `pip install torch --index-url https://download.pytorch.org/whl/cpu`
+   - Size: ~500MB
+   - Works on all Windows machines
+   - Included in the installer by default
+
+2. **voicebox-server-cuda.exe** (CUDA)
+   - Built with: `pip install torch --index-url https://download.pytorch.org/whl/cu121`
+   - Size: ~3GB
+   - Requires NVIDIA GPU + drivers
+   - Uploaded as separate GitHub Release asset
+
+### User Experience
+
+**First Launch:**
+1. User installs app (~500MB download)
+2. App starts with CPU server
+3. If NVIDIA GPU detected:
+   - Show notification: "Download CUDA support for 4-5x faster inference?"
+   - User clicks "Download"
+   - Download voicebox-server-cuda.exe from GitHub (~3GB)
+   - Save to `%APPDATA%/voicebox/binaries/`
+   - Restart server with CUDA version
+
+**Settings Panel:**
+- Toggle between CPU/CUDA modes
+- Download CUDA if not already installed
+- Show current inference backend
+
+### Build Scripts
+
+**Windows:**
+```bash
+cd backend
+
+# Build CPU only
+build_cpu.bat
+
+# Build CUDA only
+build_cuda.bat
+
+# Build both
+build_both.bat
+```
+
+**Unix (macOS/Linux):**
+```bash
+cd backend
+
+# Build CPU only
+./build_cpu.sh
+```
+
+### CI/CD Workflow
+
+**GitHub Actions (.github/workflows/release.yml):**
+
+1. Install CPU PyTorch
+2. Build CPU server → Copy to Tauri binaries
+3. Install CUDA PyTorch
+4. Build CUDA server → Save for upload
+5. Build Tauri app (bundles CPU server)
+6. Upload CUDA server as separate release asset
+
+### File Structure
+
+```
+Release Assets:
+├── Voicebox_0.1.12_x64_en-US.msi           (~500MB - includes CPU server)
+├── voicebox-server-cuda-x86_64-pc-windows-msvc.exe   (~3GB - optional download)
+└── latest.json                              (updater manifest)
+```
+
+## Implementation Details
+
+### Modified Files
+
+1. **backend/build_binary.py**
+   - Added `variant` parameter ('cpu' or 'cuda')
+   - Outputs different binary names based on variant
+
+2. **backend/build_cpu.bat** (new)
+   - Installs CPU PyTorch
+   - Builds CPU binary
+   - Restores CUDA PyTorch for dev
+
+3. **backend/build_cuda.bat** (new)
+   - Ensures CUDA PyTorch is installed
+   - Builds CUDA binary
+
+4. **.github/workflows/release.yml**
+   - Build CPU binary first (for installer)
+   - Build CUDA binary second (for upload)
+   - Upload CUDA binary as additional release asset
+   - Updated release notes to explain GPU acceleration
+
+### Future Frontend Work
+
+**TODO: Implement CUDA download in the app**
+
+Location: `tauri/src/`
+
+Features needed:
+1. GPU detection on startup
+2. Download manager for CUDA binary
+3. Server binary path switcher
+4. Settings UI for CPU/CUDA toggle
+5. Progress indicator for 3GB download
+
+API endpoints needed (already exist):
+- `/health` - Shows GPU availability
+- Server restart mechanism
+
+## Benefits
+
+✓ **Smaller installer**: ~500MB instead of 3GB
+✓ **Faster CI builds**: WiX can handle 500MB easily
+✓ **User choice**: CPU users don't download unnecessary files
+✓ **Better UX**: Optional performance upgrade for GPU users
+✓ **Cost savings**: Reduced bandwidth for users without GPUs
+
+## Testing
+
+**Test CPU build:**
+```bash
+cd backend
+python build_binary.py cpu
+./dist/voicebox-server.exe --version
+```
+
+**Test CUDA build:**
+```bash
+cd backend
+python build_binary.py cuda
+./dist/voicebox-server-cuda.exe --version
+```
+
+**Verify size:**
+```bash
+ls -lh backend/dist/
+# Should see:
+# voicebox-server.exe       ~500MB
+# voicebox-server-cuda.exe  ~3GB
+```
+
+**Test server startup:**
+```bash
+# CPU version
+./backend/dist/voicebox-server.exe
+# Check logs: Should show CPU inference
+
+# CUDA version (requires NVIDIA GPU)
+./backend/dist/voicebox-server-cuda.exe
+# Check logs: Should show CUDA inference
+```

docs/github-2gb-limit-issue.md

@@ -0,0 +1,122 @@
+# GitHub 2GB Release Asset Limit Issue
+
+## Problem
+
+The CUDA server binary upload fails in CI with:
+```
+Error: File size (2543828017) is greater than 2 GiB
+```
+
+GitHub release assets have a hard limit of 2GB per file. Our CUDA binary is ~2.5GB, which exceeds this limit.
+
+## Background
+
+The dual-server binary system (see `dual-server-binaries.md`) creates two binaries:
+- **CPU binary**: ~500MB ✅ Works fine
+- **CUDA binary**: ~2.5GB ❌ Exceeds GitHub limit
+
+## Attempted Solution: Compression
+
+We're testing 7z compression with maximum settings to see if we can squeeze the CUDA binary under 2GB.
+
+### Test Script
+
+Run `backend/test_cuda_compression.py` to test compression locally:
+
+```bash
+cd backend
+python test_cuda_compression.py
+```
+
+This will:
+1. Find the CUDA binary in `dist/`
+2. Compress it with 7z (maximum compression)
+3. Report if the compressed size fits under 2GB
+
+### Expected Compression
+
+PyTorch CUDA binaries typically compress well since they contain:
+- Repeated patterns in neural network weights
+- Debug symbols and metadata
+- Redundant CUDA libraries
+
+Estimated compression: 30-40% reduction
+- Original: ~2.5GB
+- Target: <2GB
+- Required compression: >20%
+
+## Fallback: External Hosting
+
+If compression doesn't work, we'll need to host the CUDA binary externally:
+
+### Option 1: AWS S3
+```yaml
+- name: Upload CUDA binary to S3
+  run: |
+    aws s3 cp backend/cuda-release/voicebox-server-cuda-*.exe \
+      s3://voicebox-releases/cuda-binaries/${{ github.ref_name }}/
+```
+
+### Option 2: Azure Blob Storage
+```yaml
+- name: Upload to Azure Blob
+  run: |
+    az storage blob upload \
+      --account-name voiceboxreleases \
+      --container-name cuda-binaries \
+      --file backend/cuda-release/voicebox-server-cuda-*.exe
+```
+
+### Option 3: GitHub Packages (Container Registry)
+Package as a container image, though this adds complexity for desktop app distribution.
+
+## Implementation Plan
+
+1. **Test compression locally** ← Current step
+2. **If compression works (<2GB)**:
+   - Update CI to compress before upload
+   - Update app to handle .7z downloads
+   - Add extraction step in download manager
+
+3. **If compression fails (≥2GB)**:
+   - Set up external storage (likely S3)
+   - Update CI to upload to S3
+   - Provide download URL in release notes
+   - Update app download manager to fetch from S3
+
+## CI Workflow Changes (if compression works)
+
+```yaml
+- name: Compress CUDA binary (Windows only)
+  if: matrix.platform == 'windows-latest'
+  shell: bash
+  run: |
+    cd backend/cuda-release
+    7z a -t7z -m0=lzma2 -mx=9 -mfb=64 -md=32m -ms=on \
+      voicebox-server-cuda-x86_64-pc-windows-msvc.7z \
+      voicebox-server-cuda-*.exe
+
+- name: Upload compressed CUDA server (Windows only)
+  if: matrix.platform == 'windows-latest'
+  uses: softprops/action-gh-release@v1
+  with:
+    files: backend/cuda-release/*.7z
+```
+
+## User Experience Impact
+
+### With Compression
+- Download: `voicebox-server-cuda-*.7z` (~1.5-1.8GB)
+- App extracts automatically
+- One extra step but manageable
+
+### With External Hosting
+- Download from S3/Azure URL
+- No GitHub release asset dependency
+- Potentially faster download speeds (CDN)
+
+## Status
+
+🔄 **Testing compression locally to determine viability**
+
+Results pending from local test run.

docs/r2-setup.md

@@ -0,0 +1,274 @@
+# Cloudflare R2 Setup Guide
+
+## Overview
+
+The CUDA binary (2.4GB) is hosted on Cloudflare R2 at `downloads.voicebox.sh` instead of GitHub Releases (which has a 2GB limit).
+
+## R2 Bucket Configuration
+
+✅ **Completed:**
+- Bucket created: `voicebox`
+- Custom domain configured: `downloads.voicebox.sh`
+
+## GitHub Secrets Required
+
+Add these secrets to your GitHub repository:
+
+### 1. R2_ACCESS_KEY_ID
+
+Your Cloudflare R2 API Access Key ID
+
+**How to get it:**
+1. Go to Cloudflare Dashboard → R2
+2. Click "Manage R2 API Tokens"
+3. Create API Token with "Object Read & Write" permissions
+4. Copy the "Access Key ID"
+
+**Add to GitHub:**
+```
+Repository Settings → Secrets and variables → Actions → New repository secret
+Name: R2_ACCESS_KEY_ID
+Value: <your-access-key-id>
+```
+
+### 2. R2_SECRET_ACCESS_KEY
+
+Your Cloudflare R2 Secret Access Key
+
+**How to get it:**
+- Same process as above
+- Copy the "Secret Access Key" (shown only once!)
+- Store it securely
+
+**Add to GitHub:**
+```
+Name: R2_SECRET_ACCESS_KEY
+Value: <your-secret-access-key>
+```
+
+### 3. R2_ENDPOINT
+
+Your Cloudflare R2 endpoint URL
+
+**Format:**
+```
+https://<account-id>.r2.cloudflarestorage.com
+```
+
+**How to find your account ID:**
+- Cloudflare Dashboard → R2
+- Look at the URL or bucket settings
+- Should be a string of letters/numbers
+
+**Add to GitHub:**
+```
+Name: R2_ENDPOINT
+Value: https://<your-account-id>.r2.cloudflarestorage.com
+```
+
+## Bucket Structure
+
+After CI uploads, the bucket will have this structure:
+
+```
+voicebox/
+└── cuda/
+    ├── v0.1.12/
+    │   └── voicebox-server-cuda-x86_64-pc-windows-msvc.exe
+    ├── v0.1.13/
+    │   └── voicebox-server-cuda-x86_64-pc-windows-msvc.exe
+    └── v0.2.0/
+        └── voicebox-server-cuda-x86_64-pc-windows-msvc.exe
+```
+
+## Public Access
+
+Files are uploaded with `--acl public-read`, making them accessible at:
+
+```
+https://downloads.voicebox.sh/cuda/v{VERSION}/voicebox-server-cuda-x86_64-pc-windows-msvc.exe
+```
+
+**Example:**
+```
+https://downloads.voicebox.sh/cuda/v0.1.12/voicebox-server-cuda-x86_64-pc-windows-msvc.exe
+```
+
+## Testing the Setup
+
+### Local Test Upload
+
+Before running the CI, test uploading locally:
+
+```bash
+# Set environment variables
+export AWS_ACCESS_KEY_ID="your-r2-access-key-id"
+export AWS_SECRET_ACCESS_KEY="your-r2-secret-access-key"
+export R2_ENDPOINT="https://your-account-id.r2.cloudflarestorage.com"
+
+# Install AWS CLI
+pip install awscli
+
+# Test upload (use a small test file first)
+echo "test" > test.txt
+aws s3 cp test.txt \
+  s3://voicebox/test/test.txt \
+  --endpoint-url $R2_ENDPOINT \
+  --acl public-read
+
+# Verify it's accessible
+curl https://downloads.voicebox.sh/test/test.txt
+
+# If successful, try the actual CUDA binary
+aws s3 cp backend/dist/voicebox-server-cuda.exe \
+  s3://voicebox/cuda/v0.1.12-test/voicebox-server-cuda-x86_64-pc-windows-msvc.exe \
+  --endpoint-url $R2_ENDPOINT \
+  --acl public-read
+```
+
+### Verify Upload
+
+Check if the file is accessible:
+
+```bash
+curl -I https://downloads.voicebox.sh/cuda/v0.1.12-test/voicebox-server-cuda-x86_64-pc-windows-msvc.exe
+```
+
+Should return:
+```
+HTTP/2 200
+content-length: 2545086396
+content-type: application/x-msdownload
+...
+```
+
+## CI Workflow
+
+The workflow now:
+
+1. **Builds CPU binary** → Includes in installer
+2. **Builds CUDA binary** → Uploads to R2
+3. **Release notes** → Include R2 download link
+
+### CI Steps (Windows)
+
+```yaml
+- name: Build CUDA Python server (Windows only)
+  # Builds the CUDA binary
+
+- name: Upload CUDA server to Cloudflare R2 (Windows only)
+  env:
+    AWS_ACCESS_KEY_ID: ${{ secrets.R2_ACCESS_KEY_ID }}
+    AWS_SECRET_ACCESS_KEY: ${{ secrets.R2_SECRET_ACCESS_KEY }}
+    R2_ENDPOINT: ${{ secrets.R2_ENDPOINT }}
+  run: |
+    aws s3 cp backend/cuda-release/voicebox-server-cuda-*.exe \
+      s3://voicebox/cuda/${VERSION}/... \
+      --endpoint-url $R2_ENDPOINT \
+      --acl public-read
+```
+
+## Cost Tracking
+
+Monitor your R2 usage:
+
+**Cloudflare Dashboard → R2 → voicebox → Metrics**
+
+Expected costs (per month):
+- Storage: 2.4GB × $0.015/GB = **$0.036**
+- Egress: **$0.00** (free!)
+- Class A ops: ~100 × $4.50/million = **$0.00**
+
+**Total: ~$0.04/month** (essentially free!)
+
+## Troubleshooting
+
+### Upload fails: "Access Denied"
+
+**Solution:** Check API token permissions
+- Must have "Object Read & Write" on the bucket
+- Regenerate token if needed
+
+### File not accessible at downloads.voicebox.sh
+
+**Solution:** Check custom domain configuration
+- R2 Dashboard → Bucket → Settings → Custom Domains
+- Ensure `downloads.voicebox.sh` is properly configured
+- DNS may take time to propagate
+
+### "endpoint-url" not recognized
+
+**Solution:** Make sure AWS CLI is updated
+```bash
+pip install --upgrade awscli
+```
+
+### File uploaded but wrong permissions
+
+**Solution:** Re-upload with `--acl public-read`
+```bash
+aws s3 cp ... --acl public-read
+```
+
+Or set bucket default permissions in R2 Dashboard.
+
+## Security Notes
+
+### API Token Permissions
+
+✅ **Recommended:**
+- Object Read & Write only
+- No admin permissions needed
+- Scoped to `voicebox` bucket only
+
+❌ **Avoid:**
+- Account-wide permissions
+- Account admin access
+- Worker edit permissions
+
+### Secret Rotation
+
+Rotate API tokens every 6-12 months:
+1. Create new API token
+2. Update GitHub secrets
+3. Verify CI still works
+4. Delete old token
+
+## Maintenance
+
+### Cleaning Old Versions
+
+Optional: Delete old CUDA binaries to save storage costs
+
+```bash
+# List all versions
+aws s3 ls s3://voicebox/cuda/ \
+  --endpoint-url $R2_ENDPOINT
+
+# Delete old version
+aws s3 rm s3://voicebox/cuda/v0.1.0/ \
+  --recursive \
+  --endpoint-url $R2_ENDPOINT
+```
+
+### Monitoring
+
+Set up Cloudflare notifications:
+- Storage approaching limits
+- Unusual traffic patterns
+- High operation counts
+
+## Next Steps
+
+1. ✅ Bucket configured
+2. ⏳ Add GitHub secrets (R2_ACCESS_KEY_ID, R2_SECRET_ACCESS_KEY, R2_ENDPOINT)
+3. ⏳ Test local upload
+4. ⏳ Push branch and create test release
+5. ⏳ Verify CUDA binary accessible from downloads.voicebox.sh
+6. ⏳ Implement frontend download manager
+
+---
+
+**Status**: Ready for testing
+**Cost**: ~$0.04/month
+**Bandwidth**: Free (unlimited)

landing/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@voicebox/landing",
-  "version": "0.1.11",
+  "version": "0.1.12",
   "description": "Landing page for voicebox.sh",
   "scripts": {
     "dev": "bun --bun next dev --turbo",

package.json

@@ -1,6 +1,6 @@
 {
   "name": "voicebox",
-  "version": "0.1.11",
+  "version": "0.1.12",
   "private": true,
   "workspaces": [
     "app",

tauri/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@voicebox/tauri",
   "private": true,
-  "version": "0.1.11",
+  "version": "0.1.12",
   "type": "module",
   "scripts": {
     "dev": "vite",
@@ -10,7 +10,11 @@
   },
   "dependencies": {
     "@tauri-apps/api": "^2.0.0",
-    "@tauri-apps/plugin-shell": "^2.0.0"
+    "@tauri-apps/plugin-dialog": "^2.0.0",
+    "@tauri-apps/plugin-fs": "^2.0.0",
+    "@tauri-apps/plugin-process": "^2.0.0",
+    "@tauri-apps/plugin-shell": "^2.0.0",
+    "@tauri-apps/plugin-updater": "^2.0.0"
   },
   "devDependencies": {
     "@tailwindcss/vite": "^4.1.18",

tauri/src-tauri/Cargo.lock

@@ -5041,7 +5041,7 @@ checksum = "0b928f33d975fc6ad9f86c8f283853ad26bdd5b10b7f1542aa2fa15e2289105a"
 
 [[package]]
 name = "voicebox"
-version = "0.1.11"
+version = "0.1.12"
 dependencies = [
  "base64 0.22.1",
  "core-foundation-sys",

tauri/src-tauri/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "voicebox"
-version = "0.1.11"
+version = "0.1.12"
 description = "A production-quality desktop app for Qwen3-TTS voice cloning and generation"
 authors = ["you"]
 license = ""

tauri/src-tauri/tauri.conf.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://schema.tauri.app/config/2",
   "productName": "Voicebox",
-  "version": "0.1.11",
+  "version": "0.1.12",
   "identifier": "sh.voicebox.app",
   "build": {
     "beforeDevCommand": "bun run dev",

tauri/vite.config.ts

@@ -35,5 +35,15 @@ export default defineConfig({
     minify: !process.env.TAURI_DEBUG,
     sourcemap: !!process.env.TAURI_DEBUG,
     outDir: 'dist',
+    rollupOptions: {
+      external: [
+        '@tauri-apps/api',
+        '@tauri-apps/plugin-dialog',
+        '@tauri-apps/plugin-fs',
+        '@tauri-apps/plugin-process',
+        '@tauri-apps/plugin-shell',
+        '@tauri-apps/plugin-updater',
+      ],
+    },
   },
 });

test_cuda.py

@@ -0,0 +1,77 @@
+"""Test CUDA detection in voicebox backend"""
+import sys
+import torch
+
+print("=" * 60)
+print("PyTorch CUDA Detection Test")
+print("=" * 60)
+
+# Basic torch info
+print(f"\nPyTorch version: {torch.__version__}")
+print(f"CUDA available: {torch.cuda.is_available()}")
+
+if torch.cuda.is_available():
+    print(f"CUDA version: {torch.version.cuda}")
+    print(f"GPU count: {torch.cuda.device_count()}")
+    print(f"Current GPU: {torch.cuda.current_device()}")
+    print(f"GPU name: {torch.cuda.get_device_name(0)}")
+    print(f"GPU memory: {torch.cuda.get_device_properties(0).total_memory / 1024**3:.2f} GB")
+else:
+    print("\nNo CUDA available - would run on CPU")
+
+# Test backend device selection
+print("\n" + "=" * 60)
+print("Backend Device Selection")
+print("=" * 60)
+
+# Simulate the _get_device method from pytorch_backend.py
+def _get_device() -> str:
+    """Get the best available device."""
+    if torch.cuda.is_available():
+        return "cuda"
+    elif hasattr(torch.backends, 'mps') and torch.backends.mps.is_available():
+        # MPS can have issues, use CPU for stability
+        return "cpu"
+    return "cpu"
+
+selected_device = _get_device()
+print(f"\nSelected device: {selected_device}")
+print(f"Would use dtype: {'torch.bfloat16' if selected_device != 'cpu' else 'torch.float32'}")
+
+# Test actual tensor creation on device
+print("\n" + "=" * 60)
+print("Testing Tensor Creation on Device")
+print("=" * 60)
+
+try:
+    test_tensor = torch.randn(1000, 1000).to(selected_device)
+    print(f"\n[OK] Successfully created tensor on {selected_device}")
+    print(f"  Tensor device: {test_tensor.device}")
+    print(f"  Tensor dtype: {test_tensor.dtype}")
+
+    # Test computation
+    result = test_tensor @ test_tensor.T
+    print(f"[OK] Successfully performed computation on {selected_device}")
+
+    if selected_device == "cuda":
+        print(f"\nCUDA memory allocated: {torch.cuda.memory_allocated() / 1024**2:.2f} MB")
+        print(f"CUDA memory reserved: {torch.cuda.memory_reserved() / 1024**2:.2f} MB")
+
+except Exception as e:
+    print(f"\n[ERROR] {e}")
+
+print("\n" + "=" * 60)
+print("Summary")
+print("=" * 60)
+
+if selected_device == "cuda":
+    print("\n[SUCCESS] CUDA IS WORKING!")
+    print("  The backend will use your NVIDIA GPU for inference")
+    print(f"  GPU: {torch.cuda.get_device_name(0)}")
+    print(f"  This will be significantly faster than CPU")
+else:
+    print("\n[FAIL] CUDA is not available")
+    print("  The backend will use CPU for inference")
+    print("  This will be slower than GPU")
+
+print("\n" + "=" * 60)

web/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@voicebox/web",
   "private": true,
-  "version": "0.1.11",
+  "version": "0.1.12",
   "type": "module",
   "scripts": {
     "dev": "vite",