This project demonstrates a powerful real-time translation system that automatically transcribes and translates live speech into multiple languages. Perfect for multilingual conferences, educational sessions, or international meetings.
- Real-Time Translation: Instant translation of live speech
- Multi-Language Support: Currently supports English, French, German, Spanish, and Japanese
- Single Host System: Optimized for one speaker with multiple listeners
- Language Preferences: Each listener can choose their preferred language
- High-Quality Speech Recognition: Powered by Deepgram's advanced STT
- Neural Translation: Utilizing Gemini API for accurate translations
- 🌐 LiveKit: Real-time communication infrastructure
- 🤖 LiveKit Agents: Backend processing and coordination
- 👂 Deepgram: Speech-to-text processing
- 🌍 Google Gemini AI: Neural machine translation
- ⚡ Next.js: Frontend framework
-
Room Creation & Management
- Automatic agent joining on room creation
- Dynamic host detection and mic stream subscription
-
Speech Processing Pipeline
- Real-time audio streaming
- Speech-to-text conversion via Deepgram
- Neural translation processing
-
Translation Distribution
- Language-specific routing
- Real-time caption delivery
- Multi-user synchronization
cd serverpython -m venv .venvsource .venv/bin/activatepip install -r requirements.txtcp .env.example .env- add values for keys in
.env python main.py dev
cd client/webpnpm icp .env.example .env.local- add values for keys in
.env.local pnpm dev- open a browser and navigate to
http://localhost:3000
- Single host restriction per session
- Occasional UI glitches when multiple browser windows are open
- STT performance may degrade with multiple concurrent connections
You can easily add support for additional languages by modifying the language configuration in the agent code. The system is designed to be modular and extensible.
Looking to implement a similar system for your organization? We specializes in building custom AI-powered solutions.
🔗 Contact Us for Professional Implementation
For more information about LiveKit Agents and their capabilities, visit: LiveKit Agents Documentation
This project is licensed under the Apache 2.0 License - see the LICENSE file for details.