Give Your LiveKit Voice Agent Memory with MongoDB Atlas Vector Search

A voice agent that forgets you between calls is just a phone tree with better diction. In this walkthrough I wire MongoDB Atlas Vector Search into a LiveKit voice agent in Python, covering retrieval augmented generation, personalization, and memory that carries across sessions. Sign up for LiveKit Cloud → https://cloud.livekit.io/signup?utm_source=youtube&utm_medium=video&utm_campaign=devrel&utm_content=t-NJWI2QL_E I'll walk you through a working starter kit and the five integration patterns that make it tick, so you can clone it and be talking to a memory equipped agent in under ten minutes. **What you will learn** - Running $vectorSearch inside a LiveKit function tool for RAG - Giving your agent persistent, per user memory with hybrid $rankFusion retrieval - Identifying a caller and pre loading their profile before the agent speaks - Persisting session reports on hangup with on_session_end - How Voyage AI embeddings power semantic recall in Atlas 📚 Resources 📚 Agent docs → https://docs.livekit.io/agents/?utm_source=youtube&utm_medium=video&utm_campaign=devrel&utm_content=t-NJWI2QL_E Starter kit → https://github.com/livekit-examples/mongodb-hacker-starter MongoDB Atlas Vector Search → https://www.mongodb.com/docs/vector-search/ Voyage AI embeddings → https://docs.voyageai.com/docs/embeddings 🤝 Join the Community: https://community.livekit.io/?utm_source=youtube&utm_medium=video&utm_campaign=devrel&utm_content=t-NJWI2QL_E #livekit #ai #voiceai