Home | Docs | Online Flashing | Build from Source |
简体中文 ESP-Claw is Espressif's Chat Coding AI agent framework for IoT devices. It defines device behavior through conversation and completes the full loop of sensing, decision-making, and execution locally on Espressif chips. Inspired by the OpenClaw concept and reimplemented in C, ESP-Claw is lightweight, intelligent, and continuously evolving. With just an ESP32-series chip that costs only a few dollars, you can experience what makes ESP-Claw so nimble.
Traditional IoT usually stops at connectivity: devices can connect to the network, but they cannot think; they can execute commands, but they cannot make decisions. ESP-Claw brings the Agent Runtime down onto Espressif chips, turning them from passive executors into active decision-making centers.
💬 Chat as Creation | ⚙️ Event Driven | |---|---| | IM chat + dynamic Lua Ordinary users can define device behavior without programming | Any event can trigger the Agent Loop and more Response can be as fast as milliseconds |
chat-coding-en.webm | #
event-en.webm | #
🧬 Structured Memory | 📤 MCP Communication | | Organize memories in a structured way Privacy stays off the cloud | Supports standard MCP devices Works as both Server and Client |
memory-en.webm | #
mcp-en.webm | #
🧰 Ready Out of the Box | 🧩 Component Extensibility | | Quick setup with Board Manager Supports one-click flashing | Every module can be trimmed as needed You can also add your own component integrations |
ESP-Claw now supports a wide range of development boards based on ESP32-S3, ESP32-P4, ESP32-C5, and ESP32-S31, including breadboards, M5Stack CoreS3, and many others. Supported boards in ./application/edge_agent/boards/ can be flashed online directly: configuration and flashing are done entirely in the browser, with no need to compile firmware locally or install a development environment first.
You can also build ESP-Claw locally. Please refer to the local build documentation for board adaptation, building, and flashing. Boards not listed above, as well as chips like the ESP32-P4, can also be supported through local builds and flashing.
You can find practical examples in our documentation. LLM: ESP-Claw now supports both OpenAI-style APIs and Anthropic-style APIs. It natively supports GPT models from OpenAI, Qwen models from Alibaba Cloud Bailian, Claude models from Anthropic, DeepSeek models from DeepSeek API, and also supports custom endpoints.
Tip
ESP-Claw's self-programming capability depends on models with strong tool use and instruction-following ability. We recommend gpt-5.4
, qwen3.6-plus
, claude4.6-sonnet
, deepseek-v4-pro
or models with comparable capability.
IM: ESP-Claw supports Telegram, QQ, Feishu, and WeChat, and can be extended further.
ESP-Claw is still under active development. Feel free to open an issue to report problems or request features. You can also share your ideas through our online survey (in Chinese).
Click here to view our TODO List (in Chinese) and vote for the features or issues you care about. That helps us prioritize them sooner.
If this project helps you, please consider giving it a star. ⭐⭐⭐⭐⭐
Inspired by [OpenClaw](https://github.com/openclaw/openclaw).
The implementation of Agent Loop, IM communication, and related capabilities on ESP32 also draws on MimiClaw.