macos audio capture for sample creation

app/src/App.tsx

@@ -14,7 +14,7 @@ import { UpdateNotification } from '@/components/UpdateNotification';
 import { Toaster } from '@/components/ui/toaster';
 import { ProfileList } from '@/components/VoiceProfiles/ProfileList';
 import { useModelDownloadToast } from '@/lib/hooks/useModelDownloadToast';
-import { useRestoreActiveTasks, MODEL_DISPLAY_NAMES } from '@/lib/hooks/useRestoreActiveTasks';
+import { MODEL_DISPLAY_NAMES, useRestoreActiveTasks } from '@/lib/hooks/useRestoreActiveTasks';
 import { isMacOS, isTauri, setupWindowCloseHandler, startServer } from '@/lib/tauri';
 
 // Track if server is starting to prevent duplicate starts

app/src/components/VoiceProfiles/ProfileForm.tsx

@@ -39,9 +39,11 @@ import {
 } from '@/lib/hooks/useProfiles';
 import { useTranscription } from '@/lib/hooks/useTranscription';
 import { useAudioRecording } from '@/lib/hooks/useAudioRecording';
+import { useSystemAudioCapture } from '@/lib/hooks/useSystemAudioCapture';
 import { useUIStore } from '@/stores/uiStore';
-import { Mic, Square, Upload } from 'lucide-react';
+import { Mic, Square, Upload, Monitor } from 'lucide-react';
 import { formatAudioDuration } from '@/lib/utils/audio';
+import { isTauri } from '@/lib/tauri';
 
 // Helper function to get audio duration from File
 async function getAudioDuration(file: File): Promise<number> {
@@ -101,7 +103,7 @@ export function ProfileForm() {
   const addSample = useAddSample();
   const transcribe = useTranscription();
   const { toast } = useToast();
-  const [sampleMode, setSampleMode] = useState<'upload' | 'record'>('upload');
+  const [sampleMode, setSampleMode] = useState<'upload' | 'record' | 'system'>('upload');
   const [audioDuration, setAudioDuration] = useState<number | null>(null);
   const [isValidatingAudio, setIsValidatingAudio] = useState(false);
   const isCreating = !editingProfileId;
@@ -171,6 +173,28 @@ export function ProfileForm() {
     },
   });
 
+  const {
+    isRecording: isSystemRecording,
+    duration: systemDuration,
+    error: systemRecordingError,
+    isSupported: isSystemAudioSupported,
+    startRecording: startSystemRecording,
+    stopRecording: stopSystemRecording,
+    cancelRecording: cancelSystemRecording,
+  } = useSystemAudioCapture({
+    maxDurationSeconds: 30,
+    onRecordingComplete: (blob) => {
+      const file = new File([blob], `system-audio-${Date.now()}.wav`, {
+        type: blob.type || 'audio/wav',
+      });
+      form.setValue('sampleFile', file, { shouldValidate: true });
+      toast({
+        title: 'System audio captured',
+        description: 'Audio has been captured successfully.',
+      });
+    },
+  });
+
   // Show recording errors
   useEffect(() => {
     if (recordingError) {
@@ -182,6 +206,17 @@ export function ProfileForm() {
     }
   }, [recordingError, toast]);
 
+  // Show system audio recording errors
+  useEffect(() => {
+    if (systemRecordingError) {
+      toast({
+        title: 'System audio capture error',
+        description: systemRecordingError,
+        variant: 'destructive',
+      });
+    }
+  }, [systemRecordingError, toast]);
+
   useEffect(() => {
     if (editingProfile) {
       form.reset({
@@ -234,7 +269,11 @@ export function ProfileForm() {
   }
 
   function handleCancelRecording() {
-    cancelRecording();
+    if (sampleMode === 'record') {
+      cancelRecording();
+    } else if (sampleMode === 'system') {
+      cancelSystemRecording();
+    }
     form.resetField('sampleFile');
   }
 
@@ -345,6 +384,9 @@ export function ProfileForm() {
       if (isRecording) {
         cancelRecording();
       }
+      if (isSystemRecording) {
+        cancelSystemRecording();
+      }
     }
   }
 
@@ -432,9 +474,19 @@ export function ProfileForm() {
 
                   <Tabs
                     value={sampleMode}
-                    onValueChange={(v) => setSampleMode(v as 'upload' | 'record')}
+                    onValueChange={(v) => {
+                      const newMode = v as 'upload' | 'record' | 'system';
+                      // Cancel any active recordings when switching modes
+                      if (isRecording && newMode !== 'record') {
+                        cancelRecording();
+                      }
+                      if (isSystemRecording && newMode !== 'system') {
+                        cancelSystemRecording();
+                      }
+                      setSampleMode(newMode);
+                    }}
                   >
-                    <TabsList className="grid w-full grid-cols-2">
+                    <TabsList className={`grid w-full ${isTauri() && isSystemAudioSupported ? 'grid-cols-3' : 'grid-cols-2'}`}>
                       <TabsTrigger value="upload" className="flex items-center gap-2">
                         <Upload className="h-4 w-4" />
                         Upload
@@ -443,6 +495,12 @@ export function ProfileForm() {
                         <Mic className="h-4 w-4" />
                         Record
                       </TabsTrigger>
+                      {isTauri() && isSystemAudioSupported && (
+                        <TabsTrigger value="system" className="flex items-center gap-2">
+                          <Monitor className="h-4 w-4" />
+                          System Audio
+                        </TabsTrigger>
+                      )}
                     </TabsList>
 
                     <TabsContent value="upload" className="space-y-4">
@@ -611,6 +669,103 @@ export function ProfileForm() {
                         )}
                       />
                     </TabsContent>
+
+                    {isTauri() && isSystemAudioSupported && (
+                      <TabsContent value="system" className="space-y-4">
+                        <FormField
+                          control={form.control}
+                          name="sampleFile"
+                          render={() => (
+                            <FormItem>
+                              <FormLabel>Capture System Audio</FormLabel>
+                              <FormControl>
+                                <div className="space-y-4">
+                                  {!isSystemRecording && !selectedFile && (
+                                    <div className="flex flex-col items-center gap-4 p-4 border-2 border-dashed rounded-lg">
+                                      <Button
+                                        type="button"
+                                        onClick={startSystemRecording}
+                                        size="lg"
+                                        className="flex items-center gap-2"
+                                      >
+                                        <Monitor className="h-5 w-5" />
+                                        Start Capture
+                                      </Button>
+                                      <p className="text-sm text-muted-foreground text-center">
+                                        Capture audio from your system. Maximum duration: 30 seconds.
+                                      </p>
+                                    </div>
+                                  )}
+
+                                  {isSystemRecording && (
+                                    <div className="flex flex-col items-center gap-4 p-4 border-2 border-destructive rounded-lg bg-destructive/5">
+                                      <div className="flex items-center gap-4">
+                                        <div className="flex items-center gap-2">
+                                          <div className="h-3 w-3 rounded-full bg-destructive animate-pulse" />
+                                          <span className="text-lg font-mono font-semibold">
+                                            {formatAudioDuration(systemDuration)}
+                                          </span>
+                                        </div>
+                                      </div>
+                                      <Button
+                                        type="button"
+                                        onClick={stopSystemRecording}
+                                        variant="destructive"
+                                        className="flex items-center gap-2"
+                                      >
+                                        <Square className="h-4 w-4" />
+                                        Stop Capture
+                                      </Button>
+                                      <p className="text-sm text-muted-foreground text-center">
+                                        Capturing system audio... ({formatAudioDuration(30 - systemDuration)}{' '}
+                                        remaining)
+                                      </p>
+                                    </div>
+                                  )}
+
+                                  {selectedFile && !isSystemRecording && (
+                                    <div className="flex flex-col items-center gap-4 p-4 border-2 border-primary rounded-lg bg-primary/5">
+                                      <div className="flex items-center gap-2">
+                                        <Monitor className="h-5 w-5 text-primary" />
+                                        <span className="font-medium">Capture complete</span>
+                                      </div>
+                                      <p className="text-sm text-muted-foreground text-center">
+                                        File: {selectedFile.name}
+                                      </p>
+                                      <div className="flex gap-2">
+                                        <Button
+                                          type="button"
+                                          variant="outline"
+                                          onClick={handleTranscribe}
+                                          disabled={transcribe.isPending}
+                                          className="flex items-center gap-2"
+                                        >
+                                          <Mic className="h-4 w-4" />
+                                          {transcribe.isPending ? 'Transcribing...' : 'Transcribe'}
+                                        </Button>
+                                        <Button
+                                          type="button"
+                                          variant="outline"
+                                          onClick={handleCancelRecording}
+                                          className="flex items-center gap-2"
+                                        >
+                                          Capture Again
+                                        </Button>
+                                      </div>
+                                    </div>
+                                  )}
+                                </div>
+                              </FormControl>
+                              <FormDescription>
+                                Capture audio from your system (speakers, applications). Maximum duration is 30
+                                seconds.
+                              </FormDescription>
+                              <FormMessage />
+                            </FormItem>
+                          )}
+                        />
+                      </TabsContent>
+                    )}
                   </Tabs>
 
                   <FormField

app/src/components/VoiceProfiles/SampleUpload.tsx

@@ -26,8 +26,10 @@ import { useToast } from '@/components/ui/use-toast';
 import { useAddSample, useProfile } from '@/lib/hooks/useProfiles';
 import { useTranscription } from '@/lib/hooks/useTranscription';
 import { useAudioRecording } from '@/lib/hooks/useAudioRecording';
-import { Mic, Square, Upload } from 'lucide-react';
+import { useSystemAudioCapture } from '@/lib/hooks/useSystemAudioCapture';
+import { Mic, Square, Upload, Monitor } from 'lucide-react';
 import { formatAudioDuration } from '@/lib/utils/audio';
+import { isTauri } from '@/lib/tauri';
 
 const sampleSchema = z.object({
   file: z.instanceof(File, { message: 'Please select an audio file' }),
@@ -50,7 +52,7 @@ export function SampleUpload({ profileId, open, onOpenChange }: SampleUploadProp
   const transcribe = useTranscription();
   const { data: profile } = useProfile(profileId);
   const { toast } = useToast();
-  const [mode, setMode] = useState<'upload' | 'record'>('upload');
+  const [mode, setMode] = useState<'upload' | 'record' | 'system'>('upload');
 
   const form = useForm<SampleFormValues>({
     resolver: zodResolver(sampleSchema),
@@ -83,6 +85,29 @@ export function SampleUpload({ profileId, open, onOpenChange }: SampleUploadProp
     },
   });
 
+  const {
+    isRecording: isSystemRecording,
+    duration: systemDuration,
+    error: systemRecordingError,
+    isSupported: isSystemAudioSupported,
+    startRecording: startSystemRecording,
+    stopRecording: stopSystemRecording,
+    cancelRecording: cancelSystemRecording,
+  } = useSystemAudioCapture({
+    maxDurationSeconds: 30,
+    onRecordingComplete: (blob) => {
+      // Convert blob to File object
+      const file = new File([blob], `system-audio-${Date.now()}.wav`, {
+        type: blob.type || 'audio/wav',
+      });
+      form.setValue('file', file, { shouldValidate: true });
+      toast({
+        title: 'System audio captured',
+        description: 'Audio has been captured successfully.',
+      });
+    },
+  });
+
   // Show recording errors
   useEffect(() => {
     if (recordingError) {
@@ -94,6 +119,17 @@ export function SampleUpload({ profileId, open, onOpenChange }: SampleUploadProp
     }
   }, [recordingError, toast]);
 
+  // Show system audio recording errors
+  useEffect(() => {
+    if (systemRecordingError) {
+      toast({
+        title: 'System audio capture error',
+        description: systemRecordingError,
+        variant: 'destructive',
+      });
+    }
+  }, [systemRecordingError, toast]);
+
   async function handleTranscribe() {
     const file = form.getValues('file');
     if (!file) {
@@ -154,12 +190,19 @@ export function SampleUpload({ profileId, open, onOpenChange }: SampleUploadProp
       if (isRecording) {
         cancelRecording();
       }
+      if (isSystemRecording) {
+        cancelSystemRecording();
+      }
     }
     onOpenChange(newOpen);
   }
 
   function handleCancelRecording() {
-    cancelRecording();
+    if (mode === 'record') {
+      cancelRecording();
+    } else if (mode === 'system') {
+      cancelSystemRecording();
+    }
     // Reset file field by clearing the input
     form.resetField('file');
   }
@@ -176,8 +219,13 @@ export function SampleUpload({ profileId, open, onOpenChange }: SampleUploadProp
 
         <Form {...form}>
           <form onSubmit={form.handleSubmit(onSubmit)} className="space-y-4">
-            <Tabs value={mode} onValueChange={(v) => setMode(v as 'upload' | 'record')}>
+            <Tabs
-              <TabsList className="grid w-full grid-cols-2">
+              value={mode}
+              onValueChange={(v) => setMode(v as 'upload' | 'record' | 'system')}
+            >
+              <TabsList
+                className={`grid w-full ${isTauri() && isSystemAudioSupported ? 'grid-cols-3' : 'grid-cols-2'}`}
+              >
                 <TabsTrigger value="upload" className="flex items-center gap-2">
                   <Upload className="h-4 w-4" />
                   Upload
@@ -186,6 +234,12 @@ export function SampleUpload({ profileId, open, onOpenChange }: SampleUploadProp
                   <Mic className="h-4 w-4" />
                   Record
                 </TabsTrigger>
+                {isTauri() && isSystemAudioSupported && (
+                  <TabsTrigger value="system" className="flex items-center gap-2">
+                    <Monitor className="h-4 w-4" />
+                    System Audio
+                  </TabsTrigger>
+                )}
               </TabsList>
 
               <TabsContent value="upload" className="space-y-4">
@@ -325,6 +379,103 @@ export function SampleUpload({ profileId, open, onOpenChange }: SampleUploadProp
                   )}
                 />
               </TabsContent>
+
+              {isTauri() && isSystemAudioSupported && (
+                <TabsContent value="system" className="space-y-4">
+                  <FormField
+                    control={form.control}
+                    name="file"
+                    render={() => (
+                      <FormItem>
+                        <FormLabel>Capture System Audio</FormLabel>
+                        <FormControl>
+                          <div className="space-y-4">
+                            {!isSystemRecording && !selectedFile && (
+                              <div className="flex flex-col items-center gap-4 p-6 border-2 border-dashed rounded-lg">
+                                <Button
+                                  type="button"
+                                  onClick={startSystemRecording}
+                                  size="lg"
+                                  className="flex items-center gap-2"
+                                >
+                                  <Monitor className="h-5 w-5" />
+                                  Start Capture
+                                </Button>
+                                <p className="text-sm text-muted-foreground text-center">
+                                  Capture audio from your system. Maximum duration: 30 seconds.
+                                </p>
+                              </div>
+                            )}
+
+                            {isSystemRecording && (
+                              <div className="flex flex-col items-center gap-4 p-6 border-2 border-destructive rounded-lg bg-destructive/5">
+                                <div className="flex items-center gap-4">
+                                  <div className="flex items-center gap-2">
+                                    <div className="h-3 w-3 rounded-full bg-destructive animate-pulse" />
+                                    <span className="text-lg font-mono font-semibold">
+                                      {formatAudioDuration(systemDuration)}
+                                    </span>
+                                  </div>
+                                </div>
+                                <Button
+                                  type="button"
+                                  onClick={stopSystemRecording}
+                                  variant="destructive"
+                                  className="flex items-center gap-2"
+                                >
+                                  <Square className="h-4 w-4" />
+                                  Stop Capture
+                                </Button>
+                                <p className="text-sm text-muted-foreground text-center">
+                                  Capturing system audio... ({formatAudioDuration(30 - systemDuration)}{' '}
+                                  remaining)
+                                </p>
+                              </div>
+                            )}
+
+                            {selectedFile && !isSystemRecording && mode === 'system' && (
+                              <div className="flex flex-col items-center gap-4 p-6 border-2 border-primary rounded-lg bg-primary/5">
+                                <div className="flex items-center gap-2">
+                                  <Monitor className="h-5 w-5 text-primary" />
+                                  <span className="font-medium">Capture complete</span>
+                                </div>
+                                <p className="text-sm text-muted-foreground">
+                                  File: {selectedFile.name}
+                                </p>
+                                <div className="flex gap-2">
+                                  <Button
+                                    type="button"
+                                    variant="outline"
+                                    onClick={handleTranscribe}
+                                    disabled={transcribe.isPending}
+                                    className="flex items-center gap-2"
+                                  >
+                                    <Mic className="h-4 w-4" />
+                                    {transcribe.isPending ? 'Transcribing...' : 'Transcribe'}
+                                  </Button>
+                                  <Button
+                                    type="button"
+                                    variant="outline"
+                                    onClick={handleCancelRecording}
+                                    className="flex items-center gap-2"
+                                  >
+                                    Capture Again
+                                  </Button>
+                                </div>
+                              </div>
+                            )}
+                          </div>
+                        </FormControl>
+                        <FormDescription>
+                          Capture audio from your system (applications, browser tabs, etc.). Works
+                          natively without browser dialogs.
+                        </FormDescription>
+                        <FormMessage />
+                      </FormItem>
+                    )}
+                  />
+                </TabsContent>
+              )}
             </Tabs>
 
             <FormField

app/src/lib/hooks/useSystemAudioCapture.ts

@@ -0,0 +1,165 @@
+import { useState, useRef, useCallback, useEffect } from 'react';
+import { invoke } from '@tauri-apps/api/core';
+import { isTauri } from '@/lib/tauri';
+
+interface UseSystemAudioCaptureOptions {
+  maxDurationSeconds?: number;
+  onRecordingComplete?: (blob: Blob) => void;
+}
+
+/**
+ * Hook for native system audio capture using Tauri commands.
+ * Uses ScreenCaptureKit on macOS and WASAPI loopback on Windows.
+ */
+export function useSystemAudioCapture({
+  maxDurationSeconds = 30,
+  onRecordingComplete,
+}: UseSystemAudioCaptureOptions = {}) {
+  const [isRecording, setIsRecording] = useState(false);
+  const [duration, setDuration] = useState(0);
+  const [error, setError] = useState<string | null>(null);
+  const [isSupported, setIsSupported] = useState(false);
+  const timerRef = useRef<number | null>(null);
+  const startTimeRef = useRef<number | null>(null);
+  const stopRecordingRef = useRef<(() => Promise<void>) | null>(null);
+
+  // Check if system audio capture is supported
+  useEffect(() => {
+    if (!isTauri()) {
+      setIsSupported(false);
+      return;
+    }
+
+    invoke<boolean>('is_system_audio_supported')
+      .then((supported) => {
+        setIsSupported(supported);
+      })
+      .catch(() => {
+        setIsSupported(false);
+      });
+  }, []);
+
+  const startRecording = useCallback(async () => {
+    if (!isTauri()) {
+      const errorMsg = 'System audio capture is only available in the desktop app.';
+      setError(errorMsg);
+      return;
+    }
+
+    if (!isSupported) {
+      const errorMsg = 'System audio capture is not supported on this platform.';
+      setError(errorMsg);
+      return;
+    }
+
+    try {
+      setError(null);
+      setDuration(0);
+
+      // Start native capture
+      await invoke('start_system_audio_capture', {
+        maxDurationSecs: maxDurationSeconds,
+      });
+
+      setIsRecording(true);
+      startTimeRef.current = Date.now();
+
+      // Start timer
+      timerRef.current = window.setInterval(() => {
+        if (startTimeRef.current) {
+          const elapsed = (Date.now() - startTimeRef.current) / 1000;
+          setDuration(elapsed);
+
+          // Auto-stop at max duration
+          if (elapsed >= maxDurationSeconds && stopRecordingRef.current) {
+            void stopRecordingRef.current();
+          }
+        }
+      }, 100);
+    } catch (err) {
+      const errorMessage =
+        err instanceof Error
+          ? err.message
+          : 'Failed to start system audio capture. Please check permissions.';
+      setError(errorMessage);
+      setIsRecording(false);
+    }
+  }, [maxDurationSeconds, isSupported]);
+
+  const stopRecording = useCallback(async () => {
+    if (!isRecording || !isTauri()) {
+      return;
+    }
+
+    try {
+      setIsRecording(false);
+
+      if (timerRef.current !== null) {
+        clearInterval(timerRef.current);
+        timerRef.current = null;
+      }
+
+      // Stop capture and get base64 WAV data
+      const base64Data = await invoke<string>('stop_system_audio_capture');
+
+      // Convert base64 to Blob
+      const binaryString = atob(base64Data);
+      const bytes = new Uint8Array(binaryString.length);
+      for (let i = 0; i < binaryString.length; i++) {
+        bytes[i] = binaryString.charCodeAt(i);
+      }
+
+      const blob = new Blob([bytes], { type: 'audio/wav' });
+      onRecordingComplete?.(blob);
+    } catch (err) {
+      const errorMessage =
+        err instanceof Error
+          ? err.message
+          : 'Failed to stop system audio capture.';
+      setError(errorMessage);
+    }
+  }, [isRecording, onRecordingComplete]);
+
+  // Store stopRecording in ref for use in timer
+  useEffect(() => {
+    stopRecordingRef.current = stopRecording;
+  }, [stopRecording]);
+
+  const cancelRecording = useCallback(async () => {
+    if (isRecording) {
+      await stopRecording();
+    }
+
+    setIsRecording(false);
+    setDuration(0);
+
+    if (timerRef.current !== null) {
+      clearInterval(timerRef.current);
+      timerRef.current = null;
+    }
+  }, [isRecording, stopRecording]);
+
+  // Cleanup on unmount
+  useEffect(() => {
+    return () => {
+      if (timerRef.current !== null) {
+        clearInterval(timerRef.current);
+      }
+      // Cancel recording on unmount if still recording
+      if (isRecording) {
+        void cancelRecording();
+      }
+    };
+    // biome-ignore lint/correctness/useExhaustiveDependencies: cancelRecording is stable
+  }, [isRecording]);
+
+  return {
+    isRecording,
+    duration,
+    error,
+    isSupported,
+    startRecording,
+    stopRecording,
+    cancelRecording,
+  };
+}

tauri/src-tauri/Cargo.lock

@@ -103,6 +103,24 @@ version = "0.22.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "72b3254f16251a8381aa12e40e3c4d2f0199f8c6508fbecb9d91f575e0fbb8c6"
 
+[[package]]
+name = "bindgen"
+version = "0.72.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "993776b509cfb49c750f11b8f07a46fa23e0a1386ffc01fb1e7d343efc387895"
+dependencies = [
+ "bitflags 2.10.0",
+ "cexpr",
+ "clang-sys",
+ "itertools",
+ "proc-macro2",
+ "quote",
+ "regex",
+ "rustc-hash",
+ "shlex",
+ "syn 2.0.114",
+]
+
 [[package]]
 name = "bitflags"
 version = "1.3.2"
@@ -267,6 +285,15 @@ version = "1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6d43a04d8753f35258c91f8ec639f792891f748a1edbd759cf1dcea3382ad83c"
 
+[[package]]
+name = "cexpr"
+version = "0.6.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6fac387a98bb7c37292057cffc56d62ecb629900026402633ae9160df93a8766"
+dependencies = [
+ "nom",
+]
+
 [[package]]
 name = "cfb"
 version = "0.7.3"
@@ -312,6 +339,17 @@ dependencies = [
  "windows-link 0.2.1",
 ]
 
+[[package]]
+name = "clang-sys"
+version = "1.8.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0b023947811758c97c59bf9d1c188fd619ad4718dcaa767947df1cadb14f39f4"
+dependencies = [
+ "glob",
+ "libc",
+ "libloading 0.8.9",
+]
+
 [[package]]
 name = "combine"
 version = "4.6.7"
@@ -378,6 +416,15 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "coreaudio-sys"
+version = "0.2.17"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ceec7a6067e62d6f931a2baf6f3a751f4a892595bcec1461a3c94ef9949864b6"
+dependencies = [
+ "bindgen",
+]
+
 [[package]]
 name = "cpufeatures"
 version = "0.2.17"
@@ -646,6 +693,12 @@ version = "1.0.20"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d0881ea181b1df73ff77ffaaf9c7544ecc11e82fba9b5f27b262a3c73a332555"
 
+[[package]]
+name = "either"
+version = "1.15.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "48c757948c5ede0e46177b7add2e67155f70e33c07fea8284df6576da70b3719"
+
 [[package]]
 name = "embed-resource"
 version = "3.0.6"
@@ -1212,6 +1265,12 @@ version = "0.4.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7f24254aa9a54b5c858eaee2f5bccdb46aaf0e486a595ed5fd8f86ba55232a70"
 
+[[package]]
+name = "hound"
+version = "3.5.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "62adaabb884c94955b19907d60019f4e145d091c75345379e70d1ee696f7854f"
+
 [[package]]
 name = "html5ever"
 version = "0.29.1"
@@ -1534,6 +1593,15 @@ dependencies = [
  "once_cell",
 ]
 
+[[package]]
+name = "itertools"
+version = "0.13.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "413ee7dfc52ee1a4949ceeb7dbc8a33f2d6c088194d9f922fb8318faf1f01186"
+dependencies = [
+ "either",
+]
+
 [[package]]
 name = "itoa"
 version = "1.0.17"
@@ -1666,7 +1734,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6e9ec52138abedcc58dc17a7c6c0c00a2bdb4f3427c7f63fa97fd0d859155caf"
 dependencies = [
  "gtk-sys",
- "libloading",
+ "libloading 0.7.4",
  "once_cell",
 ]
 
@@ -1686,6 +1754,16 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "libloading"
+version = "0.8.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d7c4b02199fee7c5d21a5ae7d8cfa79a6ef5bb2fc834d6e9058e89c825efdc55"
+dependencies = [
+ "cfg-if",
+ "windows-link 0.2.1",
+]
+
 [[package]]
 name = "libredox"
 version = "0.1.12"
@@ -1736,6 +1814,15 @@ version = "0.1.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c41e0c4fef86961ac6d6f8a82609f55f31b05e4fce149ac5710e439df7619ba4"
 
+[[package]]
+name = "malloc_buf"
+version = "0.0.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "62bb907fe88d54d8d9ce32a3cceab4218ed2f6b7d35617cafe9adf84e43919cb"
+dependencies = [
+ "libc",
+]
+
 [[package]]
 name = "markup5ever"
 version = "0.14.1"
@@ -1788,6 +1875,12 @@ version = "0.3.17"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6877bb514081ee2a7ff5ef9de3281f14a4dd4bceac4c09388074a6b5df8a139a"
 
+[[package]]
+name = "minimal-lexical"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
+
 [[package]]
 name = "minisign-verify"
 version = "0.2.4"
@@ -1878,12 +1971,31 @@ version = "0.1.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "72ef4a56884ca558e5ddb05a1d1e7e1bfd9a68d9ed024c21704cc98872dae1bb"
 
+[[package]]
+name = "nom"
+version = "7.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
+dependencies = [
+ "memchr",
+ "minimal-lexical",
+]
+
 [[package]]
 name = "num-conv"
 version = "0.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "cf97ec579c3c42f953ef76dbf8d55ac91fb219dde70e49aa4a6b7d74e9919050"
 
+[[package]]
+name = "num-integer"
+version = "0.1.46"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7969661fd2958a5cb096e56c8e1ad0444ac2bbcd0061bd28660485a44879858f"
+dependencies = [
+ "num-traits",
+]
+
 [[package]]
 name = "num-traits"
 version = "0.2.19"
@@ -1915,6 +2027,15 @@ dependencies = [
  "syn 2.0.114",
 ]
 
+[[package]]
+name = "objc"
+version = "0.2.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "915b1b472bc21c53464d6c8461c9d3af805ba1ef837e1cac254428f4a77177b1"
+dependencies = [
+ "malloc_buf",
+]
+
 [[package]]
 name = "objc2"
 version = "0.6.3"
@@ -3016,6 +3137,12 @@ version = "1.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "94143f37725109f92c262ed2cf5e59bce7498c01bcc1502d7b9afe439a4e9f49"
 
+[[package]]
+name = "screencapturekit"
+version = "1.5.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6ccf069cb109cf8e01ebdca0d55dfce45dbbf669e8c56ed5c62150b056d3ec9f"
+
 [[package]]
 name = "selectors"
 version = "0.24.0"
@@ -3504,7 +3631,7 @@ dependencies = [
  "tao-macros",
  "unicode-segmentation",
  "url",
- "windows",
+ "windows 0.61.3",
  "windows-core 0.61.2",
  "windows-version",
  "x11-dl",
@@ -3586,7 +3713,7 @@ dependencies = [
  "webkit2gtk",
  "webview2-com",
  "window-vibrancy",
- "windows",
+ "windows 0.61.3",
 ]
 
 [[package]]
@@ -3784,7 +3911,7 @@ dependencies = [
  "url",
  "webkit2gtk",
  "webview2-com",
- "windows",
+ "windows 0.61.3",
 ]
 
 [[package]]
@@ -3810,7 +3937,7 @@ dependencies = [
  "url",
  "webkit2gtk",
  "webview2-com",
- "windows",
+ "windows 0.61.3",
  "wry",
 ]
 
@@ -4358,6 +4485,12 @@ checksum = "0b928f33d975fc6ad9f86c8f283853ad26bdd5b10b7f1542aa2fa15e2289105a"
 name = "voicebox"
 version = "0.1.0"
 dependencies = [
+ "base64 0.22.1",
+ "core-foundation-sys",
+ "coreaudio-sys",
+ "hound",
+ "objc",
+ "screencapturekit",
  "serde",
  "serde_json",
  "tauri",
@@ -4367,6 +4500,7 @@ dependencies = [
  "tauri-plugin-shell",
  "tauri-plugin-updater",
  "tokio",
+ "wasapi",
 ]
 
 [[package]]
@@ -4408,6 +4542,19 @@ dependencies = [
  "try-lock",
 ]
 
+[[package]]
+name = "wasapi"
+version = "0.22.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7834ac561bea8a7413661fdda62180f9e054f99815269cd3572cf7e40d9c3191"
+dependencies = [
+ "log",
+ "num-integer",
+ "thiserror 2.0.18",
+ "windows 0.62.2",
+ "windows-core 0.62.2",
+]
+
 [[package]]
 name = "wasi"
 version = "0.9.0+wasi-snapshot-preview1"
@@ -4582,7 +4729,7 @@ checksum = "7130243a7a5b33c54a444e54842e6a9e133de08b5ad7b5861cd8ed9a6a5bc96a"
 dependencies = [
  "webview2-com-macros",
  "webview2-com-sys",
- "windows",
+ "windows 0.61.3",
  "windows-core 0.61.2",
  "windows-implement",
  "windows-interface",
@@ -4606,7 +4753,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "381336cfffd772377d291702245447a5251a2ffa5bad679c99e61bc48bacbf9c"
 dependencies = [
  "thiserror 2.0.18",
- "windows",
+ "windows 0.61.3",
  "windows-core 0.61.2",
 ]
 
@@ -4662,11 +4809,23 @@ version = "0.61.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9babd3a767a4c1aef6900409f85f5d53ce2544ccdfaa86dad48c91782c6d6893"
 dependencies = [
- "windows-collections",
+ "windows-collections 0.2.0",
  "windows-core 0.61.2",
- "windows-future",
+ "windows-future 0.2.1",
  "windows-link 0.1.3",
- "windows-numerics",
+ "windows-numerics 0.2.0",
+]
+
+[[package]]
+name = "windows"
+version = "0.62.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "527fadee13e0c05939a6a05d5bd6eec6cd2e3dbd648b9f8e447c6518133d8580"
+dependencies = [
+ "windows-collections 0.3.2",
+ "windows-core 0.62.2",
+ "windows-future 0.3.2",
+ "windows-numerics 0.3.1",
 ]
 
 [[package]]
@@ -4678,6 +4837,15 @@ dependencies = [
  "windows-core 0.61.2",
 ]
 
+[[package]]
+name = "windows-collections"
+version = "0.3.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "23b2d95af1a8a14a3c7367e1ed4fc9c20e0a26e79551b1454d72583c97cc6610"
+dependencies = [
+ "windows-core 0.62.2",
+]
+
 [[package]]
 name = "windows-core"
 version = "0.61.2"
@@ -4712,7 +4880,18 @@ checksum = "fc6a41e98427b19fe4b73c550f060b59fa592d7d686537eebf9385621bfbad8e"
 dependencies = [
  "windows-core 0.61.2",
  "windows-link 0.1.3",
- "windows-threading",
+ "windows-threading 0.1.0",
+]
+
+[[package]]
+name = "windows-future"
+version = "0.3.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e1d6f90251fe18a279739e78025bd6ddc52a7e22f921070ccdc67dde84c605cb"
+dependencies = [
+ "windows-core 0.62.2",
+ "windows-link 0.2.1",
+ "windows-threading 0.2.1",
 ]
 
 [[package]]
@@ -4759,6 +4938,16 @@ dependencies = [
  "windows-link 0.1.3",
 ]
 
+[[package]]
+name = "windows-numerics"
+version = "0.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6e2e40844ac143cdb44aead537bbf727de9b044e107a0f1220392177d15b0f26"
+dependencies = [
+ "windows-core 0.62.2",
+ "windows-link 0.2.1",
+]
+
 [[package]]
 name = "windows-result"
 version = "0.3.4"
@@ -4897,6 +5086,15 @@ dependencies = [
  "windows-link 0.1.3",
 ]
 
+[[package]]
+name = "windows-threading"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3949bd5b99cafdf1c7ca86b43ca564028dfe27d66958f2470940f73d86d75b37"
+dependencies = [
+ "windows-link 0.2.1",
+]
+
 [[package]]
 name = "windows-version"
 version = "0.1.7"
@@ -5123,7 +5321,7 @@ dependencies = [
  "webkit2gtk",
  "webkit2gtk-sys",
  "webview2-com",
- "windows",
+ "windows 0.61.3",
  "windows-core 0.61.2",
  "windows-version",
  "x11-dl",

tauri/src-tauri/Cargo.toml

@@ -20,6 +20,17 @@ tauri-plugin-shell = "2.0"
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
 tokio = { version = "1", features = ["full"] }
+hound = "3.5"
+base64 = "0.22"
+
+[target.'cfg(target_os = "macos")'.dependencies]
+screencapturekit = { version = "1", features = ["async"] }
+coreaudio-sys = "0.2"
+objc = "0.2"
+core-foundation-sys = "0.8"
+
+[target.'cfg(target_os = "windows")'.dependencies]
+wasapi = "0.22"
 
 [target.'cfg(not(any(target_os = "android", target_os = "ios")))'.dependencies]
 tauri-plugin-updater = "2.0"

tauri/src-tauri/Info.plist

@@ -8,5 +8,7 @@
 	<string>voicebox</string>
 	<key>NSMicrophoneUsageDescription</key>
 	<string>voicebox needs microphone access to record voice samples for voice cloning.</string>
+	<key>NSScreenCaptureUsageDescription</key>
+	<string>Voicebox needs screen capture access to record system audio for voice samples.</string>
 </dict>
 </plist>

tauri/src-tauri/build.rs

@@ -1,6 +1,24 @@
 use std::process::Command;
 
 fn main() {
+    // Link Swift runtime libraries for screencapturekit crate
+    #[cfg(target_os = "macos")]
+    {
+        // Add Swift runtime library paths to RPATH
+        println!("cargo:rustc-link-arg=-Wl,-rpath,/usr/lib/swift");
+        println!("cargo:rustc-link-arg=-L/usr/lib/swift");
+        
+        // Also try Xcode's Swift libraries
+        if let Ok(output) = Command::new("xcode-select").arg("-p").output() {
+            if output.status.success() {
+                let xcode_path = String::from_utf8_lossy(&output.stdout).trim().to_string();
+                let swift_lib_path = format!("{}/Toolchains/XcodeDefault.xctoolchain/usr/lib/swift/macosx", xcode_path);
+                println!("cargo:rustc-link-arg=-Wl,-rpath,{}", swift_lib_path);
+                println!("cargo:rustc-link-arg=-L{}", swift_lib_path);
+            }
+        }
+    }
+
     // Compile macOS Liquid Glass icon
     #[cfg(target_os = "macos")]
     {

tauri/src-tauri/src/audio_capture/macos.rs

@@ -0,0 +1,213 @@
+use crate::audio_capture::AudioCaptureState;
+use base64::{engine::general_purpose, Engine as _};
+use hound::{WavSpec, WavWriter};
+use screencapturekit::{
+    cm::CMSampleBuffer,
+    shareable_content::SCShareableContent,
+    stream::{
+        configuration::SCStreamConfiguration,
+        content_filter::SCContentFilter,
+        output_trait::SCStreamOutputTrait,
+        output_type::SCStreamOutputType,
+        sc_stream::SCStream,
+    },
+};
+use std::io::Cursor;
+use std::sync::{Arc, Mutex};
+use tokio::sync::mpsc;
+
+pub async fn start_capture(
+    state: &AudioCaptureState,
+    max_duration_secs: u32,
+) -> Result<(), String> {
+    // Reset previous samples
+    state.reset();
+
+    // Get shareable content
+    let content = SCShareableContent::get()
+        .map_err(|e| format!("Failed to get shareable content: {}", e))?;
+
+    // Get first display
+    let displays = content.displays();
+    if displays.is_empty() {
+        return Err("No displays available".to_string());
+    }
+    let display = &displays[0];
+
+    // Create content filter for desktop audio
+    let filter = SCContentFilter::create()
+        .with_display(display)
+        .with_excluding_windows(&[])
+        .build();
+
+    // Create stream configuration - audio only
+    let mut config = SCStreamConfiguration::default();
+    config.set_captures_audio(true);
+    config.set_excludes_current_process_audio(false);
+    config.set_sample_rate(48000); // Use i32 directly
+    config.set_channel_count(2); // Use i32 directly
+
+    // Create stream using builder
+    let (tx, mut rx) = mpsc::channel::<()>(1);
+    *state.stop_tx.lock().unwrap() = Some(tx);
+
+    let samples = state.samples.clone();
+    let sample_rate = state.sample_rate.clone();
+    let channels = state.channels.clone();
+
+    // Set sample rate and channels
+    *sample_rate.lock().unwrap() = 48000;
+    *channels.lock().unwrap() = 2;
+
+    // Create output handler struct
+    struct AudioHandler {
+        samples: Arc<Mutex<Vec<f32>>>,
+    }
+
+    impl SCStreamOutputTrait for AudioHandler {
+        fn did_output_sample_buffer(
+            &self,
+            sample: CMSampleBuffer,
+            _type: SCStreamOutputType,
+        ) {
+            if _type == SCStreamOutputType::Audio {
+                if let Ok(audio_samples) = extract_audio_samples(sample) {
+                    let mut samples_guard = self.samples.lock().unwrap();
+                    samples_guard.extend_from_slice(&audio_samples);
+                }
+            }
+        }
+    }
+
+    let handler = AudioHandler {
+        samples: samples.clone(),
+    };
+
+    // Create stream
+    let mut stream = SCStream::new(&filter, &config);
+    
+    // Add output handler for audio (order: handler, then output_type)
+    stream.add_output_handler(handler, SCStreamOutputType::Audio);
+
+    // Store stream reference
+    *state.stream.lock().unwrap() = Some(stream.clone());
+
+    stream.start_capture().map_err(|e| format!("Failed to start capture: {}", e))?;
+
+    // Spawn task to stop after max duration
+    let stream_clone = stream.clone();
+    tokio::spawn(async move {
+        tokio::select! {
+            _ = tokio::time::sleep(tokio::time::Duration::from_secs(max_duration_secs as u64)) => {
+                // Timeout reached
+            }
+            _ = rx.recv() => {
+                // Manual stop
+            }
+        }
+        let _ = stream_clone.stop_capture();
+    });
+
+    Ok(())
+}
+
+pub async fn stop_capture(state: &AudioCaptureState) -> Result<String, String> {
+    // Signal stop
+    if let Some(tx) = state.stop_tx.lock().unwrap().take() {
+        let _ = tx.send(());
+    }
+
+    // Stop stream if still active
+    if let Some(stream) = state.stream.lock().unwrap().take() {
+        let _ = stream.stop_capture();
+    }
+
+    // Wait a bit for capture to stop
+    tokio::time::sleep(tokio::time::Duration::from_millis(500)).await;
+
+    // Get samples
+    let samples = state.samples.lock().unwrap().clone();
+    let sample_rate = *state.sample_rate.lock().unwrap();
+    let channels = *state.channels.lock().unwrap();
+
+    if samples.is_empty() {
+        return Err("No audio samples captured".to_string());
+    }
+
+    // Convert to WAV
+    let wav_data = samples_to_wav(&samples, sample_rate, channels)?;
+    
+    // Encode to base64
+    let base64_data = general_purpose::STANDARD.encode(&wav_data);
+    
+    Ok(base64_data)
+}
+
+pub fn is_supported() -> bool {
+    // ScreenCaptureKit requires macOS 12.3+
+    // Check if we're on a supported version
+    #[cfg(target_os = "macos")]
+    {
+        // Basic check - ScreenCaptureKit should be available on macOS 12.3+
+        true
+    }
+    #[cfg(not(target_os = "macos"))]
+    {
+        false
+    }
+}
+
+fn extract_audio_samples(sample_buffer: CMSampleBuffer) -> Result<Vec<f32>, String> {
+    // Use the crate's built-in method to get audio buffer list
+    let audio_buffer_list = sample_buffer
+        .audio_buffer_list()
+        .ok_or_else(|| "Failed to get audio buffer list".to_string())?;
+
+    let mut samples = Vec::new();
+
+    // Iterate through audio buffers
+    for buffer in audio_buffer_list.iter() {
+        // Get raw bytes and interpret as f32 samples
+        let data_bytes = buffer.data();
+        let num_samples = data_bytes.len() / std::mem::size_of::<f32>();
+        
+        if num_samples > 0 {
+            unsafe {
+                // Interpret bytes as f32 samples
+                let data_ptr = data_bytes.as_ptr() as *const f32;
+                let data = std::slice::from_raw_parts(data_ptr, num_samples);
+                samples.extend_from_slice(data);
+            }
+        }
+    }
+
+    Ok(samples)
+}
+
+fn samples_to_wav(samples: &[f32], sample_rate: u32, channels: u16) -> Result<Vec<u8>, String> {
+    let mut buffer = Vec::new();
+    let cursor = Cursor::new(&mut buffer);
+    
+    let spec = WavSpec {
+        channels,
+        sample_rate,
+        bits_per_sample: 16,
+        sample_format: hound::SampleFormat::Int,
+    };
+
+    let mut writer = WavWriter::new(cursor, spec)
+        .map_err(|e| format!("Failed to create WAV writer: {}", e))?;
+
+    // Convert f32 samples to i16
+    for sample in samples {
+        let clamped = sample.clamp(-1.0, 1.0);
+        let i16_sample = (clamped * 32767.0) as i16;
+        writer.write_sample(i16_sample)
+            .map_err(|e| format!("Failed to write sample: {}", e))?;
+    }
+
+    writer.finalize()
+        .map_err(|e| format!("Failed to finalize WAV: {}", e))?;
+
+    Ok(buffer)
+}

tauri/src-tauri/src/audio_capture/mod.rs

@@ -0,0 +1,40 @@
+#[cfg(target_os = "macos")]
+mod macos;
+#[cfg(target_os = "windows")]
+mod windows;
+
+#[cfg(target_os = "macos")]
+pub use macos::*;
+#[cfg(target_os = "windows")]
+pub use windows::*;
+
+use std::sync::{Arc, Mutex};
+
+#[cfg(target_os = "macos")]
+use screencapturekit::stream::sc_stream::SCStream;
+
+pub struct AudioCaptureState {
+    pub samples: Arc<Mutex<Vec<f32>>>,
+    pub sample_rate: Arc<Mutex<u32>>,
+    pub channels: Arc<Mutex<u16>>,
+    pub stop_tx: Arc<Mutex<Option<tokio::sync::mpsc::Sender<()>>>>,
+    #[cfg(target_os = "macos")]
+    pub stream: Arc<Mutex<Option<SCStream>>>,
+}
+
+impl AudioCaptureState {
+    pub fn new() -> Self {
+        Self {
+            samples: Arc::new(Mutex::new(Vec::new())),
+            sample_rate: Arc::new(Mutex::new(44100)),
+            channels: Arc::new(Mutex::new(2)),
+            stop_tx: Arc::new(Mutex::new(None)),
+            #[cfg(target_os = "macos")]
+            stream: Arc::new(Mutex::new(None)),
+        }
+    }
+
+    pub fn reset(&self) {
+        *self.samples.lock().unwrap() = Vec::new();
+    }
+}

tauri/src-tauri/src/audio_capture/windows.rs

@@ -0,0 +1,181 @@
+use crate::audio_capture::AudioCaptureState;
+use base64::{engine::general_purpose, Engine as _};
+use hound::{WavSpec, WavWriter};
+use std::io::Cursor;
+use std::sync::{Arc, Mutex};
+use tokio::sync::mpsc;
+use wasapi::*;
+
+pub async fn start_capture(
+    state: &AudioCaptureState,
+    max_duration_secs: u32,
+) -> Result<(), String> {
+    // Reset previous samples
+    state.reset();
+
+    // Get default audio render device for loopback
+    let device = DeviceEnumerator::new()
+        .map_err(|e| format!("Failed to create device enumerator: {}", e))?
+        .get_default_audio_endpoint(&Direction::Render)
+        .map_err(|e| format!("Failed to get default render device: {}", e))?;
+
+    // Create audio client for loopback capture
+    let audio_client = device
+        .get_iaudioclient()
+        .map_err(|e| format!("Failed to get audio client: {}", e))?;
+
+    // Get mix format
+    let mix_format = audio_client
+        .get_mixformat()
+        .map_err(|e| format!("Failed to get mix format: {}", e))?;
+
+    // Set sample rate and channels
+    *state.sample_rate.lock().unwrap() = mix_format.get_samples_per_sec();
+    *state.channels.lock().unwrap() = mix_format.get_nchannels();
+
+    // Initialize audio client for loopback
+    audio_client
+        .initialize_client(
+            &mix_format,
+            0, // Buffer duration (0 = default)
+            &Direction::Capture,
+            ShareMode::Shared,
+            true, // Loopback mode
+        )
+        .map_err(|e| format!("Failed to initialize audio client: {}", e))?;
+
+    // Get capture client
+    let capture_client = audio_client
+        .get_audiocaptureclient()
+        .map_err(|e| format!("Failed to get capture client: {}", e))?;
+
+    // Start capture
+    audio_client
+        .start_stream()
+        .map_err(|e| format!("Failed to start stream: {}", e))?;
+
+    let samples = state.samples.clone();
+    let stop_tx = state.stop_tx.clone();
+    let (tx, mut rx) = mpsc::channel::<()>(1);
+    *stop_tx.lock().unwrap() = Some(tx);
+
+    // Spawn capture task - move audio_client and capture_client into the task
+    tokio::spawn(async move {
+        loop {
+            tokio::select! {
+                _ = rx.recv() => {
+                    break;
+                }
+                _ = tokio::time::sleep(tokio::time::Duration::from_millis(10)) => {
+                    // Try to get available data
+                    match capture_client.get_available_samples() {
+                        Ok(available) => {
+                            if available > 0 {
+                                match capture_client.get_buffer::<f32>() {
+                                    Ok((data, flags)) => {
+                                        if flags.contains(&StreamFlags::SILENT) {
+                                            // Silent buffer, skip
+                                            capture_client.release_buffer(available).ok();
+                                            continue;
+                                        }
+                                        
+                                        // Convert samples to f32 and store
+                                        let mut samples_guard = samples.lock().unwrap();
+                                        samples_guard.extend_from_slice(data);
+                                        capture_client.release_buffer(available).ok();
+                                    }
+                                    Err(e) => {
+                                        eprintln!("Error getting buffer: {}", e);
+                                    }
+                                }
+                            }
+                        }
+                        Err(e) => {
+                            eprintln!("Error getting available samples: {}", e);
+                        }
+                    }
+                }
+            }
+        }
+        
+        // Stop the stream when done
+        audio_client.stop_stream().ok();
+    });
+
+    // Spawn timeout task
+    let stop_tx_clone = state.stop_tx.clone();
+    tokio::spawn(async move {
+        tokio::time::sleep(tokio::time::Duration::from_secs(max_duration_secs as u64)).await;
+        if let Some(tx) = stop_tx_clone.lock().unwrap().take() {
+            let _ = tx.send(());
+        }
+    });
+
+    Ok(())
+}
+
+pub async fn stop_capture(state: &AudioCaptureState) -> Result<String, String> {
+    // Signal stop
+    if let Some(tx) = state.stop_tx.lock().unwrap().take() {
+        let _ = tx.send(());
+    }
+
+    // Wait a bit for capture to stop
+    tokio::time::sleep(tokio::time::Duration::from_millis(500)).await;
+
+    // Get samples
+    let samples = state.samples.lock().unwrap().clone();
+    let sample_rate = *state.sample_rate.lock().unwrap();
+    let channels = *state.channels.lock().unwrap();
+
+    if samples.is_empty() {
+        return Err("No audio samples captured".to_string());
+    }
+
+    // Convert to WAV
+    let wav_data = samples_to_wav(&samples, sample_rate, channels)?;
+    
+    // Encode to base64
+    let base64_data = general_purpose::STANDARD.encode(&wav_data);
+    
+    Ok(base64_data)
+}
+
+pub fn is_supported() -> bool {
+    #[cfg(target_os = "windows")]
+    {
+        true
+    }
+    #[cfg(not(target_os = "windows"))]
+    {
+        false
+    }
+}
+
+fn samples_to_wav(samples: &[f32], sample_rate: u32, channels: u16) -> Result<Vec<u8>, String> {
+    let mut buffer = Vec::new();
+    let cursor = Cursor::new(&mut buffer);
+    
+    let spec = WavSpec {
+        channels,
+        sample_rate,
+        bits_per_sample: 16,
+        sample_format: hound::SampleFormat::Int,
+    };
+
+    let mut writer = WavWriter::new(cursor, spec)
+        .map_err(|e| format!("Failed to create WAV writer: {}", e))?;
+
+    // Convert f32 samples to i16
+    for sample in samples {
+        let clamped = sample.clamp(-1.0, 1.0);
+        let i16_sample = (clamped * 32767.0) as i16;
+        writer.write_sample(i16_sample)
+            .map_err(|e| format!("Failed to write sample: {}", e))?;
+    }
+
+    writer.finalize()
+        .map_err(|e| format!("Failed to finalize WAV: {}", e))?;
+
+    Ok(buffer)
+}

tauri/src-tauri/src/main.rs

@@ -1,6 +1,8 @@
 // Prevents additional console window on Windows in release, DO NOT REMOVE!!
 #![cfg_attr(not(debug_assertions), windows_subsystem = "windows")]
 
+mod audio_capture;
+
 use std::sync::Mutex;
 use tauri::{command, State, Manager, WindowEvent, Emitter, Listener};
 use tauri_plugin_shell::ShellExt;
@@ -165,6 +167,26 @@ async fn stop_server(state: State<'_, ServerState>) -> Result<(), String> {
     Ok(())
 }
 
+#[command]
+async fn start_system_audio_capture(
+    state: State<'_, audio_capture::AudioCaptureState>,
+    max_duration_secs: u32,
+) -> Result<(), String> {
+    audio_capture::start_capture(&state, max_duration_secs).await
+}
+
+#[command]
+async fn stop_system_audio_capture(
+    state: State<'_, audio_capture::AudioCaptureState>,
+) -> Result<String, String> {
+    audio_capture::stop_capture(&state).await
+}
+
+#[command]
+fn is_system_audio_supported() -> bool {
+    audio_capture::is_supported()
+}
+
 #[cfg_attr(mobile, tauri::mobile_entry_point)]
 pub fn run() {
     tauri::Builder::default()
@@ -174,6 +196,7 @@ pub fn run() {
         .manage(ServerState {
             child: Mutex::new(None),
         })
+        .manage(audio_capture::AudioCaptureState::new())
         .setup(|app| {
             #[cfg(desktop)]
             app.handle().plugin(tauri_plugin_updater::Builder::new().build())?;
@@ -190,7 +213,13 @@ pub fn run() {
             }
             Ok(())
         })
-        .invoke_handler(tauri::generate_handler![start_server, stop_server])
+        .invoke_handler(tauri::generate_handler![
+            start_server,
+            stop_server,
+            start_system_audio_capture,
+            stop_system_audio_capture,
+            is_system_audio_supported
+        ])
         .on_window_event(|window, event| {
             if let WindowEvent::CloseRequested { api, .. } = event {
                 // Prevent automatic close