Frequently Asked Questions (FAQ) about ros2-mcp-server

Q: What is ros2-mcp-server? A: ros2-mcp-server is a Python-based server that integrates the Model Context Protocol (MCP) with ROS 2. It allows AI assistants to control robots via ROS 2 topics by processing commands through FastMCP and publishing geometry_msgs/Twist messages to control robot movement.

Q: What is MCP? A: MCP stands for Model Context Protocol. It’s an open protocol that standardizes how applications provide context to LLMs, enabling them to interact with external data sources and tools.

Q: What are the prerequisites for using ros2-mcp-server? A: You need ROS 2 (Humble distribution), Python 3.10, the uv Python package manager, and the following Python packages: rclpy, fastmcp, and numpy.

Q: How do I install ros2-mcp-server? A: First, clone the repository. Then, create a uv environment using Python 3.10, activate the environment, and install the dependencies using uv pip install -e ..

Q: How do I configure ros2-mcp-server for Claude? A: You can configure it for Claude Desktop or Cline (VSCode Extension) by adding a new MCP server with the provided JSON configuration, ensuring you replace /path/to/ros2-mcp-server with the correct path to your repository.

Q: What kind of commands can I send to the robot using ros2-mcp-server? A: You can send commands to move the robot forward, backward, left, or right, specifying the linear and angular velocities, and the duration of the movement. An example is: “Please make the robot move forward at 0.2 m/s for 5 seconds.”

Q: How can I test ros2-mcp-server? A: You can test it with a simulator like Gazebo or with a real robot. Monitor the /cmd_vel topic using ros2 topic echo /cmd_vel to verify the commands being sent.

Q: What should I do if I encounter ROS 2 logging errors? A: Ensure that the ROS_LOG_DIR environment variable is set to a writable directory (e.g., /tmp).

Q: What Python version should I use? A: You should use Python 3.10, as ROS 2 Humble is built for this version.

Q: What are the limitations of ros2-mcp-server? A: Currently, it only supports the /cmd_vel topic with Twist messages and basic movement commands. Extending it for other topics or more complex behaviors requires additional implementation.

Q: Where can I find the project’s license information? A: The project is licensed under the MIT License. Also, it uses FastMCP, which is licensed under the Apache License 2.0.

Q: How does ros2-mcp-server integrate with the UBOS platform? A: ros2-mcp-server enhances AI agents within the UBOS platform, allowing for orchestrated agent management, enterprise data connectivity, custom agent creation, and the development of sophisticated multi-agent robotic systems.