robot-mcp-server

Name: robot-mcp-server
Author: showkeyjar

@showkeyjar

Repository

About robot-mcp-server

mcp server for robot and automations

#robotics #automation #mcp-server

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "robot-mcp-server": {
      "command": "python",
      "args": [
        "-m",
        "src.main",
        "--backend",
        "mock",
        "--transport",
        "stdio"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is robot-mcp-server?

It is an MCP server for controlling Unitree robots and DJI Tello drones, designed for LLMs, agent systems, and automation platforms. It supports both mock and hardware backends so workflows can be validated without real devices.

How to use robot-mcp-server?

Install Python 3.10+, create a virtual environment, and install dependencies from requirements.txt. Start with .\run.ps1 or python -m src.main --backend mock --transport stdio. Use environment variables (e.g., ROBOT_MCP_BACKEND, ROBOT_MCP_TRANSPORT) or CLI flags to configure backend, transport, host, and port. Prefer mock mode during development; switch to hardware only after validation.

Key features of robot-mcp-server

Mock and hardware backends for safe development and testing.
Unified configuration via environment variables and CLI overrides.
Health check, status inspection, and emergency-stop tools.
Built-in routine templates for common robot workflows.
Sequence executor to run multiple routines in one request.
Separate tool sets for Unitree robots and DJI Tello drones.

Use cases of robot-mcp-server

Validate robot control workflows using simulated mock mode before hardware deployment.
Execute patrol, inspection, and observation routines on Unitree robots or DJI drones.
Chain multiple robot actions (e.g., move, pause, takeoff) into a single sequence request.
Integrate robot control into LLM-driven agents or automation platforms via MCP tools.
Rapidly prototype and test new behaviors using built-in routine templates.

FAQ from robot-mcp-server

What is the difference between mock and hardware backends?

Mock mode simulates robot responses without physical hardware, allowing safe development and testing. Hardware mode connects to real Unitree robots or DJI Tello drones.

Which robots are supported?

The server supports Unitree robots (e.g., Go, B2 series) and DJI Tello drones.

How do I install the Unitree SDK?

If using real Unitree hardware, run pip install git+https://github.com/unitreerobotics/unitree_sdk2_python.git after installing core dependencies.

What transports are available?

The server supports stdio and streamable-http transport, configurable via the ROBOT_MCP_TRANSPORT environment variable or the --transport CLI flag.

What are the known limitations?

Unitree hardware control is a lightweight SDK integration layer intended as a starting point; tests mainly cover mock mode, and real hardware behavior requires on-site integration testing. Exact streamable-http behavior depends on the installed mcp package version.