active-call is a standalone Rust crate designed for building AI Voice Agents. It was originally decoupled from rustpbx to provide a dedicated, high-performance library for voice agent integration.

This crate handles the complex low-level details of voice communication, making it easy to connect LLM/NLP engines with telephony platforms. It manages:

• ASR (Automatic Speech Recognition): Handling speech-to-text transitions and events.
• TTS (Text-to-Speech): Managing real-time audio synthesis and playback.
• SIP & WebRTC: Bridging traditional telephony and modern web communication protocols.
• Voice Details: Managing audio buffers, codecs (like Opus), and real-time streaming.

By abstracting these technical hurdles, active-call allows developers to focus on building intelligent dialogue systems rather than worrying about the underlying voice infrastructure.

• Standalone Crate: Decoupled from the main PBX logic for better modularity.
• LLM/NLP Friendly: Designed to easily feed ASR results into LLMs and stream TTS responses back to the caller.
• Multi-Protocol Support: Supports SIP, WebRTC, and raw WebSocket audio streams.
• Real-time Performance: Built with Rust for low-latency audio processing.

For detailed information on REST endpoints and WebSocket protocols, please refer to the API Documentation.

docker pull ghcr.io/restsend/active-call:latest

Copy the example config and customize it:

cp active-call.example.toml config.toml

docker run -d \
  --name active-call \
  -p 8080:8080 \
  -p 13050:13050/udp \
  -p 20000-20100:20000-20100/udp \
  -v $(pwd)/config.toml:/app/config.toml:ro \
  -v $(pwd)/config:/app/config \
  ghcr.io/restsend/active-call:latest

If you have API keys, save them in an .env file:

TENCENT_APPID=your_app_id
TENCENT_SECRET_ID=your_secret_id
TENCENT_SECRET_KEY=your_secret_key
DASHSCOPE_API_KEY=your_dashscope_api_key

And mount it in the container like :

  -v $(pwd)/.env:/app/.env \

Use small range 20000-20100 for local development, bigger range like 20000-30000, or host network for production.

This project is licensed under the MIT License.