search

livekit / agents

Build real-time multimodal AI applications 🤖🎙️📹
https://docs.livekit.io/agents
Apache License 2.0
645 stars 103 forks source link

How to integrate Langchain RAG pipeline #375

Open Meet-Fefar opened 3 weeks ago

Meet-Fefar commented 3 weeks ago

I am looking for a solution to integrate langchain so they can talk to specific Vector DB. Also build special flow if user ask for specific things to LLM.

sandeepone commented 2 weeks ago

Did you look into openAI tools? You need to create retrieval tool and pass it to likevkit agent function context. Check livekit function calling example.

KarMa-31 commented 2 weeks ago

I am also working on something similar where I am trying to use function calling feature in livekit. Did you find anything on integrating Langchain here?

sandeepone commented 2 weeks ago

We're generating context using a custom mini model instead of LangChain. Function calling will interact with our microservice, which handles embeddings, vector DB, context truncation (max 2000 tokens), and business logic. We're currently in pre-production, with lots of work ahead.