An open-source Chrome extension that transforms your browser into a real-time simultaneous interpreter ("cabin translation") for English audio and video into natural Vietnamese. Powered by the free Gemini Live API for instant streaming, or 100% offline local AI with VietAI envit5 translation and VieNeu-TTS neural voice synthesis.
Leveraging the free Gemini Live API (gemini-3.5-live-translate-preview) via Google AI Studio, this extension streams tab audio directly through WebSockets without needing any server:
- Free Tier Ready: Connect with a free API key from Google AI Studio — no credit card required.
- Zero-Setup (Client-Only): Real-time AudioWorklet captures 16kHz PCM audio directly in Chrome and speaks Vietnamese translation with near-zero latency.
- Natural Dubbing: Balanced background audio with natural AI voices (
Aoede,Kore,Charon,Puck,Fenrir). - Ideal for: Livestreams, university lectures (Coursera, edX), tech keynotes, podcasts, and global news.
| Mode | Architecture | Highlights |
|---|---|---|
| ✨ Gemini Cloud API (Default) | Client-only (Chrome MV3) | Real-time Live interpretation, bilingual web translation (Gemini Flash Lite), and subtitle TTS. Zero server installation. |
| 💻 Local Server (Self-Hosted) | Python FastAPI + CTranslate2 | 100% offline & private: High-accuracy translation with VietAI envit5 and natural 48kHz speech synthesis with VieNeu-TTS v3 Turbo (or Piper TTS). Supports segmented dubbing. |
Toggle any feature instantly from the popup menu:
- 🎙️ Gemini Live Audio: Real-time simultaneous cabin interpretation.
- 🎬 Video Subtitles: Floating bilingual subtitle overlay on YouTube and HTML5 videos.
- 🔊 Voice Dubbing (TTS): Automatic spoken translation for subtitles.
- 🌐 Web Page Translation: Bilingual translation for text and articles on scroll.
- ⚡ Gemini Flash Lite: Toggle text translation to optimize quota usage.
- ✂️ Segmented Dubbing: Segment-by-segment dubbing in local server mode.
- Download
translator-0.4.0-chrome.zipfrom Releases. - Unzip the archive.
- Open Chrome and navigate to
chrome://extensions. - Enable Developer mode (top right) → Click Load unpacked → Select the unzipped folder (or the
extension/directory). - Open the extension popup, paste your Gemini API Key, and click Save & Apply.
- On YouTube, click the Microphone (Live) button on the player controls to start instant cabin interpretation!
git clone https://github.com/dangdtr/translator-oss.git
cd translator-oss
npm install
npm run buildLoad the generated .output/chrome-mv3 folder into chrome://extensions.
Run 100% privately on your CPU using VietAI envit5 for translation and VieNeu-TTS for natural Vietnamese speech.
Requires Python 3.10+ and uv:
./setup.sh # Installs VietAI envit5 and VieNeu-TTS models
./start.sh # Starts FastAPI local server on port 8765In the popup menu, switch to the 💻 Local Server tab and click Save & Apply.
npm run dev # Dev server with HMR
npm run build # Production build
npm run zip # Package release zip
npm run typecheck # TypeScript type verification
npm test # Run 52 automated unit testsThis repository is licensed under a Dual-License model:
- Chrome Extension (
src/,extension/): MIT License - Backend Local Server (
server/): GNU GPL v3.0 (due to Piper TTS engine integration) - AI Models & Third-Party Assets: See THIRD_PARTY.md for individual model licenses (OpenRAIL, Apache-2.0, CC BY 4.0).