tinqiao-oss/clawtouch-mcp
by Various
ClawTouch MCP server — exposes a real USB-HID keyboard/mouse (Raspberry Pi Pico 2) as Model Context Protocol tools for any LLM agent. MIT.
MCP
tinqiao-oss/clawtouch-mcp
Added 7 June 2026
Overview
ClawTouch MCP server exposes a real USB-HID keyboard and mouse connected via a Raspberry Pi Pico 2 as Model Context Protocol tools. This allows any LLM agent to programmatically send keystrokes and mouse actions through a physical hardware interface. The project is open source under the MIT license and written in Python.
Best for
Best for
Developers who need a reliable physical HID bridge for LLM-driven automation on locked-down or embedded systems
Use cases
- Automating GUI interactions on a machine that cannot use virtual input devices
- Testing or debugging software that requires physical HID input
- Enabling an LLM agent to control a separate computer or kiosk via USB
How to use
Install
pip install clawtouch-mcp Tested with
Claude Desktop, Cline, Continue, Cursor, OpenClaw, Hermes Agent, ChatGPT Desktop, Codex CLI
Example client config
{\n "mcpServers": {\n "clawtouch": {\n "command": "clawtouch-mcp",\n "args": ["--port", "COM7", "--screen", "1920x1080"]\n }\n }\n} Notes
ClawTouch MCP server exposes a real USB-HID keyboard and mouse connected via a Raspberry Pi Pico 2 as Model Context Protocol tools. This allows any LLM agent to programmatically send keystrokes and mouse actions through a physical hardware interface. The project is open source under the MIT license and written in Python.
2 stars on GitHub. Last updated 2026-06-07. Licensed MIT.
Use cases
- Automating GUI interactions on a machine that cannot use virtual input devices
- Testing or debugging software that requires physical HID input
- Enabling an LLM agent to control a separate computer or kiosk via USB
Pros
- Provides genuine hardware-level input, bypassing OS virtual device restrictions
- Simple MCP interface makes it easy to integrate with any LLM agent
- Low-cost hardware requirement (Raspberry Pi Pico 2)
Cons
- Requires additional hardware and wiring, not a pure software solution
- Limited to the capabilities of the Pico 2 (e.g., no joystick or gamepad support)
- Very small community (2 stars) with minimal documentation or support
Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.
Pros
- Provides genuine hardware-level input, bypassing OS virtual device restrictions
- Simple MCP interface makes it easy to integrate with any LLM agent
- Low-cost hardware requirement (Raspberry Pi Pico 2)
Cons
- Requires additional hardware and wiring, not a pure software solution
- Limited to the capabilities of the Pico 2 (e.g., no joystick or gamepad support)
- Very small community (2 stars) with minimal documentation or support
Open-source & AI alternatives
Swap-in tools that solve the same job. Weigh the trade-offs before you commit.
Pairs with
Other entries in the index that connect to this one. Click through to see the chain.
Get the free Developer’s Field Guide
A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.
Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.
