広告掲載
MCP.so
ログイン

Auto MCP Server

@snick-m

リポジトリ

基本情報

カテゴリ

その他

ライセンス

BSD-3-Clause license

トランスポート

stdio

作者

snick-m

Auto MCP Server について

This package provides necessary tooling to launch an MCP server for the ROS network. It provides support for introspection and manual config.

Auto 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": {
    "auto_mcp": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://localhost:5000/sse"
      ]
    }
  }
}

Auto 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。

上のボタンで一時的なサンドボックスでこのサーバーを実行し、ツール一覧を取得できます。

概要

What is Auto MCP?

Auto MCP is a drop-in MCP (Model Context Protocol) server for any ROS‑based system. It uses introspection to detect all active ROS topics and their interfaces at startup, then creates an SSE‑over‑HTTP transport MCP server on port 5000, enabling an LLM to interact with a robot.

How to use Auto MCP?

Clone the repository into a colcon workspace’s src directory, build with colcon build --packages-select auto_mcp, then run ros2 run auto_mcp server. This launches the node, performs introspection, and starts the MCP server bound to 0.0.0.0:5000; the SSE endpoint is at http://localhost:5000/sse. For Claude Desktop, use the npx mcp-remote proxy with the URL set to http://localhost:5000/sse in the MCP configuration. If the mcp Python package is not available via rosdep, install it manually with pip install mcp.

Key features of Auto MCP

  • Drop‑in MCP server for any ROS system
  • Automatic topic and interface detection via introspection
  • SSE over HTTP transport on port 5000
  • Simple integration with any MCP‑compatible LLM client
  • Works with Claude Desktop using a local proxy workaround

Use cases of Auto MCP

  • Give a ROS‑based robot a reasoning and planning layer via an LLM
  • Connect a robotic system to a conversational AI assistant
  • Enable natural‑language control of ROS topics and services

よくある質問

1

What is MCP?

The Model Context Protocol (MCP) is a protocol developed by Anthropic (and now supported by OpenAI) that allows LLMs to interact with various tools. Auto MCP uses this protocol to expose ROS topics as MCP tools.

2

Why should I use Auto MCP instead of writing a custom ROS–LLM bridge?

Auto MCP automates the bridge: it introspects the ROS system at runtime, discovers topics and their interfaces, and creates a standard MCP server – no manual tool definition or topic mapping is needed.

3

What are the dependencies and runtime requirements?

You need a ROS 2 workspace (e.g., Foxy or later) with colcon. The `mcp` Python package (MCP SDK) must be installed via `pip install mcp` until the rosdep pull request is merged.

4

How does the transport work?

The server uses SSE (Server‑Sent Events) over HTTP on `0.0.0.0:5000`. The SSE endpoint is accessible at `http://localhost:5000/sse` from the local machine.

5

Can I use Auto MCP with Claude Desktop?

Yes, but Claude Desktop does not natively support SSE MCP servers. Use a local proxy (e.g., `npx mcp-remote`) and configure the MCP command to point to `http://localhost:5000/sse` as shown in the README.

コメント

「その他」の他のコンテンツ

P

PickFu MCP

PickFu MCP is a Model Context Protocol (MCP) server that connects AI assistants, developers, marketers, product managers, and AI agents to PickFu's consumer research platform. It enables users to create surveys, conduct

Activepieces

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

23.2K

Awesome Mlops

visenger

A curated list of references for MLOps

14.0K

Glasswarp

Glasswarp

See and control a real Windows PC you own — from any MCP client, locally or remotely. Observe (UIA + screenshots), click/type/drag/scroll, launch apps, owner Live View. BYOH: your machine, your key.

Klarix Intelligence Engine

Klarix

B2B competitive intelligence, live sales battlecards, 5-axis ICP prospect fit scoring, and tech stack teardowns over MCP and REST with cited public-web evidence.

広告掲載