Skip to content

Voice-to-text dictation app with local (Nvidia Parakeet/Whisper) and cloud models (BYOK). Privacy-first and available cross-platform.

feat: bundled in ffmpeg

Gabriel Stein274ebca4d7e5
diff --git a/LOCAL_WHISPER_SETUP.md b/LOCAL_WHISPER_SETUP.md
index af14b15cb..4b9334d33 100644
--- a/LOCAL_WHISPER_SETUP.md
+++ b/LOCAL_WHISPER_SETUP.md
@@ -72,16 +72,25 @@ Your OpenWispr app has been successfully refactored to support **local Whisper m
- ✅ **Fast & Efficient**: Minimal local resource usage
- ✅ **Always Latest Model**: OpenAI's most advanced model
-## 📦 Requirements (Already Installed)
+## 📦 Requirements
-The following are already installed and configured:
+### ✅ Bundled Dependencies (No Installation Required)
-- ✅ **Python 3**: Installed via Homebrew
-- ✅ **FFmpeg**: Installed via Homebrew
-- ✅ **OpenAI Whisper**: Installed via pip3
-- ✅ **Bridge Script**: Created and configured
+The following are now **bundled with the app** and require no user installation:
+
+- ✅ **FFmpeg**: Bundled with the app (no system installation needed)
+- ✅ **Python Bridge Script**: Included in the app bundle
- ✅ **App Integration**: Fully implemented
+### 🔧 Optional Dependencies (For Local Whisper)
+
+For **local Whisper processing**, users still need:
+
+- ✅ **Python 3**: Installed via Homebrew or system package manager
+- ✅ **OpenAI Whisper**: Installed via `pip3 install openai-whisper`
+
+**Note**: FFmpeg is now bundled, so users no longer need to install it separately!
+
## 🧪 Testing Your Setup
### Test 1: Check Installation Status
diff --git a/electron-builder.json b/electron-builder.json
index dcef0cf7b..0dce09158 100644
--- a/electron-builder.json
+++ b/electron-builder.json
@@ -22,10 +22,12 @@
],
"asarUnpack": [
"assets/**/*",
- "whisper_bridge.py"
+ "whisper_bridge.py",
+ "node_modules/ffmpeg-static/**/*"
],
"extraResources": [
- "assets/**/*"
+ "assets/**/*",
+ "node_modules/ffmpeg-static/**/*"
],
"mac": {
"category": "public.app-category.productivity",
diff --git a/package.json b/package.json
index a73de1119..9e0e9d560 100644
--- a/package.json
+++ b/package.json
@@ -45,6 +45,7 @@
"eslint": "^9.25.0",
"eslint-plugin-react-hooks": "^5.2.0",
"eslint-plugin-react-refresh": "^0.4.19",
+ "ffmpeg-static": "^5.2.0",
"globals": "^16.0.0",
"postcss": "^8.5.6",
"tailwindcss": "^4.1.10",
diff --git a/preload.js b/preload.js
index 06f8937d3..96a631073 100644
--- a/preload.js
+++ b/preload.js
@@ -46,6 +46,8 @@ contextBridge.exposeInMainWorld("electronAPI", {
deleteWhisperModel: (modelName) =>
ipcRenderer.invoke("delete-whisper-model", modelName),
cancelWhisperDownload: () => ipcRenderer.invoke("cancel-whisper-download"),
+ checkFFmpegAvailability: () =>
+ ipcRenderer.invoke("check-ffmpeg-availability"),
// Window control functions
windowMinimize: () => ipcRenderer.invoke("window-minimize"),
diff --git a/setup.js b/setup.js
index 4420fb392..54303b978 100644
--- a/setup.js
+++ b/setup.js
@@ -1,7 +1,7 @@
-const fs = require("fs")
-const path = require("path")
+const fs = require("fs");
+const path = require("path");
-console.log("Setting up OpenWispr...")
+console.log("Setting up OpenWispr...");
const envTemplate = `# OpenAI API Configuration
# Get your API key from: https://platform.openai.com/api-keys
@@ -16,13 +16,13 @@ WHISPER_MODEL=whisper-1
LANGUAGE=
# Optional: Debug mode (set to 'true' to enable verbose logging)
-DEBUG=false`
+DEBUG=false`;
if (!fs.existsSync(".env")) {
- fs.writeFileSync(".env", envTemplate)
- console.log("✅ Created .env file template")
+ fs.writeFileSync(".env", envTemplate);
+ console.log("✅ Created .env file template");
} else {
- console.log("⚠️ .env file already exists")
+ console.log("⚠️ .env file already exists");
}
console.log(`
@@ -38,8 +38,13 @@ Features:
- Space bar to start/stop recording
- ESC to close the app
- Automatic text pasting at cursor location
+- FFmpeg bundled (no separate installation needed)
Note: Make sure you have the necessary system permissions for:
- Microphone access
- Accessibility permissions (for text pasting)
-`)
+
+For local Whisper processing, you'll also need:
+- Python 3 (install via Homebrew: brew install python3)
+- OpenAI Whisper (install via: pip3 install openai-whisper)
+`);
diff --git a/src/helpers/ipcHandlers.js b/src/helpers/ipcHandlers.js
index fb5ab12a8..e7e71cc27 100644
--- a/src/helpers/ipcHandlers.js
+++ b/src/helpers/ipcHandlers.js
@@ -217,6 +217,10 @@ class IPCHandlers {
return this.whisperManager.cancelDownload();
});
+ ipcMain.handle("check-ffmpeg-availability", async (event) => {
+ return this.whisperManager.checkFFmpegAvailability();
+ });
+
// Utility handlers
ipcMain.handle("cleanup-app", async (event) => {
try {
diff --git a/src/helpers/whisper.js b/src/helpers/whisper.js
index 4c8528bdb..4aed071e3 100644
--- a/src/helpers/whisper.js
+++ b/src/helpers/whisper.js
@@ -272,6 +272,59 @@ class WhisperManager {
}
}
+ async checkFFmpegAvailability() {
+ try {
+ const pythonCmd = await this.findPythonExecutable();
+ const whisperScriptPath = this.getWhisperScriptPath();
+
+ const result = await new Promise((resolve) => {
+ const checkProcess = spawn(pythonCmd, [
+ whisperScriptPath,
+ "--mode",
+ "check-ffmpeg",
+ ]);
+
+ let output = "";
+ let stderr = "";
+
+ checkProcess.stdout.on("data", (data) => {
+ output += data.toString();
+ });
+
+ checkProcess.stderr.on("data", (data) => {
+ stderr += data.toString();
+ });
+
+ checkProcess.on("close", (code) => {
+ if (code === 0) {
+ try {
+ const result = JSON.parse(output);
+ resolve(result);
+ } catch (parseError) {
+ resolve({
+ available: false,
+ error: "Failed to parse FFmpeg check result",
+ });
+ }
+ } else {
+ resolve({
+ available: false,
+ error: stderr || "FFmpeg check failed",
+ });
+ }
+ });
+
+ checkProcess.on("error", (error) => {
+ resolve({ available: false, error: error.message });
+ });
+ });
+
+ return result;
+ } catch (error) {
+ return { available: false, error: error.message };
+ }
+ }
+
async installWhisper() {
try {
console.log("🔧 Starting automatic Whisper installation...");
diff --git a/whisper_bridge.py b/whisper_bridge.py
index 82418c27a..2dfc1b1b2 100755
--- a/whisper_bridge.py
+++ b/whisper_bridge.py
@@ -16,6 +16,46 @@ import time
import requests
import gc
+def get_ffmpeg_path():
+ """Get path to bundled FFmpeg executable"""
+ if getattr(sys, 'frozen', False):
+ # Running in Electron app (packed)
+ base_path = sys._MEIPASS
+ else:
+ # Running in development
+ base_path = os.path.dirname(os.path.abspath(__file__))
+
+ # Platform-specific FFmpeg paths
+ if sys.platform == "darwin": # macOS
+ ffmpeg_path = os.path.join(base_path, "..", "node_modules", "ffmpeg-static", "ffmpeg")
+ elif sys.platform == "win32": # Windows
+ ffmpeg_path = os.path.join(base_path, "..", "node_modules", "ffmpeg-static", "ffmpeg.exe")
+ else: # Linux
+ ffmpeg_path = os.path.join(base_path, "..", "node_modules", "ffmpeg-static", "ffmpeg")
+
+ # Make sure the path exists
+ if os.path.exists(ffmpeg_path):
+ return ffmpeg_path
+
+ # Fallback: try to find system FFmpeg
+ import subprocess
+ try:
+ result = subprocess.run(['which', 'ffmpeg'], capture_output=True, text=True)
+ if result.returncode == 0:
+ return result.stdout.strip()
+ except:
+ pass
+
+ return None
+
+# Set FFmpeg path for Whisper
+ffmpeg_path = get_ffmpeg_path()
+if ffmpeg_path:
+ os.environ["FFMPEG_BINARY"] = ffmpeg_path
+ print(f"Using FFmpeg at: {ffmpeg_path}", file=sys.stderr)
+else:
+ print("Warning: FFmpeg not found, Whisper may not work properly", file=sys.stderr)
+
# Global model cache to avoid reloading
_model_cache = {}
@@ -328,10 +368,50 @@ def transcribe_audio(audio_path, model_name="base", language=None):
"success": False
}
+def check_ffmpeg():
+ """Check if FFmpeg is available and working"""
+ try:
+ # Test FFmpeg by trying to get its version
+ import subprocess
+ result = subprocess.run([ffmpeg_path or "ffmpeg", "-version"],
+ capture_output=True, text=True, timeout=10)
+
+ if result.returncode == 0:
+ return {
+ "available": True,
+ "path": ffmpeg_path or "ffmpeg",
+ "version": result.stdout.split('\n')[0] if result.stdout else "Unknown",
+ "success": True
+ }
+ else:
+ return {
+ "available": False,
+ "error": f"FFmpeg returned code {result.returncode}",
+ "success": False
+ }
+ except subprocess.TimeoutExpired:
+ return {
+ "available": False,
+ "error": "FFmpeg check timed out",
+ "success": False
+ }
+ except FileNotFoundError:
+ return {
+ "available": False,
+ "error": "FFmpeg not found in PATH",
+ "success": False
+ }
+ except Exception as e:
+ return {
+ "available": False,
+ "error": str(e),
+ "success": False
+ }
+
def main():
parser = argparse.ArgumentParser(description="Whisper Bridge for OpenWispr")
parser.add_argument("--mode", default="transcribe",
- choices=["transcribe", "download", "check", "list", "delete"],
+ choices=["transcribe", "download", "check", "list", "delete", "check-ffmpeg"],
help="Operation mode (default: transcribe)")
parser.add_argument("audio_file", nargs="?", help="Path to audio file to transcribe")
parser.add_argument("--model", default="base",
@@ -361,6 +441,10 @@ def main():
result = delete_model(args.model)
print(json.dumps(result))
return
+ elif args.mode == "check-ffmpeg":
+ result = check_ffmpeg()
+ print(json.dumps(result))
+ return
elif args.mode == "transcribe":
# Check if audio file exists
if not args.audio_file: