Langchain MCP Server

@rectalogic

リポジトリ

基本情報

カテゴリ

AI とエージェント

ライセンス

MIT

トランスポート

stdio

作者

rectalogic

Langchain MCP Server について

Model Context Protocol tool support for LangChain

Langchain MCP Server の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "langchain-mcp": {
      "command": "uv",
      "args": [
        "run",
        "tests/demo.py",
        "Read and summarize the file ./LICENSE"
      ]
    }
  }
}

Langchain MCP Server のツール

ツール一覧はまだありません

この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is langchain-mcp?

langchain-mcp provides Model Context Protocol (MCP) tool calling support within LangChain. It enables LangChain applications to use tools exposed by any MCP server by wrapping them as LangChain BaseTool instances. This is for developers building LangChain-based agents or chains that need to call external tools via the MCP standard.

How to use langchain-mcp?

Create a langchain_mcp.MCPToolkit by passing it an mcp.ClientSession, then call await toolkit.initialize() followed by toolkit.get_tools() to retrieve a list of LangChain BaseTools. The toolkit handles the MCP client‑server lifecycle and tool discovery.

Key features of langchain-mcp

Integrates MCP tool calling into LangChain workflows
Exposes an async MCPToolkit class for session management
Automatically converts MCP tools into BaseTool objects
Works with any MCP server that communicates over a ClientSession

Use cases of langchain-mcp

Adding filesystem, database, or API tools from MCP servers to a LangChain agent
Combining MCP tools with LangChain’s chat models and chains in a unified pipeline

よくある質問

Is this the official LangChain MCP integration?

No. LangChain now has a more official implementation called `langchain-mcp-adapters`. This repository is a community alternative.

What does langchain-mcp require?

It requires Python and can be installed via PyPI as `langchain-mcp`. The demo also uses the `uv` tool for running.

How do I run the included demo?

Set the `GROQ_API_KEY` environment variable and execute `uv run tests/demo.py "Read and summarize the file ./LICENSE"` (or equivalent) against an MCP filesystem server.

「AI とエージェント」の他のコンテンツ

Agent Traffic Lab

ATL doesn’t just help agents find tools — it gets the job done. It finds and chooses the right AI provider, tool, or MCP server, executes the task, and switches to a fallback when a provider fails. It can optimize for lo

AON — Agent Offer Network

AON

AON MCP connects AI agents to real, buyable products and services with live pricing and tracked links—enabling builders to earn developer revenue share from attributed clicks and conversions.

Legion MCP

faulkj

MCP-native LLM councils for debates, juries, blind panels, voting, refinement, and custom multi-model deliberation.

Ranla

Kurtis Tryber

AI growth agent for software businesses. Ranla ships pages that rank on Google, gets your product named in AI answers, replies on Reddit and sends lifecycle email. Ask how growth is going, have Ranla draft the work and a

fhirHydrant