agents
A framework for building realtime voice AI agents 🤖🎙️📹
What is agents?
A framework for building realtime voice AI agents 🤖🎙️📹
Overview of agents
What it is
The Agent Framework is designed for building realtime, programmable participants that run on servers. It is used to create conversational, multi-modal voice agents that can see, hear, and understand.
Key features
- Flexible integrations: A comprehensive ecosystem to mix and match the right STT, LLM, TTS, and Realtime API.
- Integrated job scheduling: Built-in task scheduling and distribution with dispatch APIs to connect end users to agents.
- Extensive WebRTC clients: Build client applications using LiveKit's open-source SDK ecosystem, supporting all major platforms.
- Telephony integration: Works with LiveKit's telephony stack, allowing agents to make or receive phone calls.
- Exchange data with clients: Use RPCs and other Data APIs to exchange data with clients.
- Semantic turn detection: Uses a transformer model to detect when a user is done with their turn, reducing interruptions.
- MCP support: Native support for MCP; integrate tools from MCP servers with one line of code.
- Builtin test framework: Write tests and use judges to ensure agent performance.
- Open-source: Fully open-source, allowing you to run the entire stack on your own servers, including LiveKit server.
Use cases
- Building realtime voice agents for conversational applications.
- Multi-agent handoff scenarios (e.g., an intro agent gathering information then handing off to a story agent).
- Outbound phone calling agents.
- Integrating tools from MCP servers.
- Testing agent behavior with automated tests.
Getting started
Install the core Agents library with plugins for popular model providers:
pip install "livekit-agents[openai,deepgram,cartesia]"
For more detailed setup and examples, see the repository README.
License and pricing
Apache-2.0 license. Open source.
Based on the project's public README and repository metadata.