English | 简体中文
| Page | What's in it |
|---|---|
| Quick start | Prerequisites, Docker and TURN ports, connecting a client, wiring your backend behind a tool call, fully-local setup |
| Capabilities | What the runtime does today, supported endpoints, AI integrations |
| Bring your own agent | The five ways to keep the intelligence in your own stack, including the HTTP agent endpoint and the llm.Client interface |
| Agent runtime | Optional built-in agent: plugins, skills, RAG, document ingestion |
| Providers | STT / LLM / TTS options, Grok speech-to-speech, MiniMax, local VibeVoice |
| Configuration | Full annotated config.toml reference |
| Protocol | WHIP signaling, DataChannel events, auth |
| Architecture | Media flow, why Go, package layout |
| Roadmap / TODO | Checklist of what is not built yet, grouped by area |
Deploying on AWS EC2: infrastructure/aws/ec2. Contributing: CONTRIBUTING.md. Security: SECURITY.md.