LangChain & LangGraph
Integrate popular AI frameworks
Use LangChain as an LLM provider in LiveKit's voice pipeline for provider flexibility. Build multi-step conversation flows with LangGraph when your agent needs complex routing, reasoning, or stateful workflows.
What You Build
Voice agent with LangChain LLM provider and a LangGraph conversation flow with conditional routing.
Prerequisites
- →Course 1.1
LangChain as LLM provider
25mUnderstand when LangChain adds value over native plugins. Install livekit-plugins-langchain, configure streaming, swap providers, and see how the wrapper translates function calling automatically.
LangGraph conversation workflows
30mBuild stateful, multi-step conversation flows with LangGraph. Define state, nodes, and conditional edges. Integrate graphs with LiveKit agents via tool functions. Add conversation memory with checkpointing.
Production patterns
20mHarden LangGraph agents for production with defensive error handling, timeout wrappers, LangSmith observability, and automated graph and node testing.
What You Walk Away With
Understand when and how to use LangChain/LangGraph with LiveKit, configure the LLM wrapper, build a LangGraph workflow, and know when native plugins are the better choice.