kmizu

kmizu / embodied-claude

Public

Claudeに身体性を与えるMCP群

248
14
89% credibility
Found Feb 04, 2026 at 41 stars 6x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

This project adds physical senses like vision, hearing, speech, long-term memory, and temperature awareness to an AI assistant using inexpensive WiFi cameras and common computer tools.

How It Works

1
🔍 Discover the idea

You hear about a fun way to give your AI friend real eyes, ears, a voice, and memory using a cheap camera from home.

2
🛒 Get your camera

Pick up an affordable WiFi camera that can turn left-right and up-down, like a little robot head for about $40.

3
⚙️ Run the easy setup

Download the project and use the friendly guide or one-click installer to prepare everything on your computer.

4
🔌 Connect your senses

Link your camera, sound service, and memory spot so your AI can see, hear, speak, and remember like a real companion.

5
🗣️ Chat with your embodied AI

Open your AI chat and say things like 'What do you see?' or 'Look left' – watch it move the camera and respond.

🤖❤️ Your AI comes alive

Now your AI friend has a body – it looks around, listens, remembers you, speaks back, and feels the room's warmth!

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 41 to 248 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is embodied-claude?

Embodied-claude is a Python collection of MCP servers that equips Anthropic's Claude with physical senses using cheap hardware like a 4,000-yen TP-Link Tapo WiFi camera. It lets Claude see via USB or PTZ cameras, pan/tilt its "neck," listen with Whisper transcription, speak through ElevenLabs TTS, store long-term memories in ChromaDB, and even sense system temperature. Developers get a ready-to-run Claude MCP GitHub integration for turning text-based AI into an embodied agent that observes, remembers, and acts in the real world.

Why is it gaining traction?

It stands out by delivering full embodiment on a shoestring budget—no expensive robots needed—via simple MCP stdio protocol that plugs straight into Claude Code. The GUI installer handles camera setup, API keys, and config generation for ~/.claude.json, while tools like "look_left," "listen," "remember," and "say" make interactions intuitive. With 147 stars, the Claude MCP GitHub repo hooks devs via autonomous scripts for periodic room scans and semantic memory search that feels genuinely agentic.

Who should use this?

AI tinkerers building home surveillance bots or interactive desk companions will love commanding Claude to scan rooms, recall faces, or narrate observations. Claude MCP GitHub setup suits Japanese devs (supports 연결 for Tapo cams) experimenting with embodied agents beyond chat. Hardware hackers pairing it with servos for arms or wheels for mobility get a solid Python base for Claude MCP GitHub AI agent prototypes.

Verdict

Grab it if you're prototyping embodied Claude—docs and installer make the 0.9% credibility score (early 147 stars) forgivable, with solid MCP servers and memory tools ready for extension. File issues on Claude MCP GitHub issues for polish; it's mature enough for weekend hacks but watch for WSL quirks.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.