You can now build production-ready AI and agent applications faster using the new LangChain and LangGraph integration for Azure Cosmos DB. This Python package enables you to directly use Cosmos DB for vector and hybrid search, semantic caching, chat history storage, checkpointing, and long-term memory, all within familiar LangChain and LangGraph workflows.
You can store and retrieve embeddings, combine full-text and vector search for more relevant results, cache LLM responses to reduce cost and latency, persist conversational history, and manage agent state with checkpointing for reliable execution. You no longer need to stitch together multiple services or databases for retrieval, memory, and orchestration. With Cosmos DB’s global distribution, scalability, and enterprise-grade reliability, you can build copilots, multi-agent systems, and knowledge assistants that operate efficiently at scale while keeping your operational and agentic data all in one place.
Learn more.