LiveKit Agents
Open-source framework for realtime multimodal voice and video AI agents.
About
LiveKit Agents is an open-source framework for building realtime voice, video, and data AI agents on top of LiveKit's WebRTC infrastructure. Agents join LiveKit rooms as full realtime participants with sub-100ms media transport, chaining speech-to-text, LLM reasoning, and text-to-speech plugins from providers like Deepgram, OpenAI, ElevenLabs, and Cartesia.
The Python and Node.js SDKs expose Agent and AgentSession abstractions with function tooling, voice activity detection, and turn handling. Agents can be deployed to LiveKit Cloud with managed hosting, phone numbers, transcripts, and observability, or self-hosted anywhere. An in-browser Agent Builder supports no-code prototyping.
What it can do
- Realtime voice and video agents over WebRTC
- Pluggable STT, LLM, TTS, and VAD providers
- Function tooling for live API calls during conversations
- Sub-100ms media transport with interruption handling
- Managed deployment on LiveKit Cloud or self-hosted
Setup
Install with pip install livekit-agents plus provider plugins, and connect to LiveKit Cloud or a self-hosted server.
Prerequisites
Python 3.10+ or Node.js, provider API keys (STT, LLM, TTS), and a LiveKit server or cloud account.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 108; Trending score 84/100 from researcher assessment.
★ 5,000 GitHub stars
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.