langchain-mcp-adapters and langgraph-mcp
These are complements: the adapter library provides the foundational integration layer for connecting MCP servers to LangChain, while the agent template demonstrates how to build autonomous agents that utilize those MCP-connected tools within LangGraph's agentic framework.
About langchain-mcp-adapters
langchain-ai/langchain-mcp-adapters
LangChain 🔌 MCP
Bridges Anthropic's Model Context Protocol (MCP) servers with LangChain/LangGraph by converting MCP tools into LangChain-compatible tools, supporting multiple transport types (stdio, HTTP, streamable HTTP, SSE) and server connections simultaneously. Provides a `MultiServerMCPClient` for managing multiple MCP servers and runtime authentication headers, enabling agentic workflows to access distributed tool ecosystems.
About langgraph-mcp
prsdm/langgraph-mcp
LangGraph agent template with MCP.
Implements automatic tool discovery across multiple MCP servers (web search, YouTube transcripts, math, weather) that LangGraph agents can dynamically invoke without hardcoding integrations. The architecture runs MCP servers separately and connects via a client-server pattern, enabling modular composition of external capabilities. Supports multi-server orchestration where agents automatically discover and select available tools based on task requirements.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work