Making AI Assistants Cheap Again
中文 | English
RT-Claw is an OpenClaw-inspired intelligent assistant for embedded devices. Multi-RTOS support via OSAL. Build swarm intelligence with networked nodes. ESP32-S3 WiFi support adapted from MimiClaw.
Deploy your own AI assistant on hardware that costs just one dollar — seamlessly integrated into your daily workflow, efficiently bridging the digital and physical worlds.
Getting Started · Usage · Architecture · Porting · Tuning · Contributing
| Feature | Description | Status |
|---|---|---|
| LLM Chat Engine | Interactive conversation with LLM API over HTTP | Done |
| Tool Use | LLM-driven function calling to interact with hardware and services; 30+ built-in tools | Done |
| Skills | Composable multi-tool workflows; AI can create, persist, and execute skills that orchestrate multiple tools | Done |
| Conversation Memory | Short-term RAM ring buffer + long-term NVS Flash persistent storage; AI can save/delete/list memories | Done |
| Swarm Intelligence | Node discovery, heartbeat, capability bitmap, remote tool invocation across nodes | Done |
| Scheduled Tasks | Timer-driven task execution and periodic automation; AI can create/list/remove tasks | Done |
| Chat-first Shell | UART REPL with insert-mode editing, tab completion, UTF-8; direct input goes to AI, /commands for system | Done |
| OSAL | Write once, run on FreeRTOS and RT-Thread with zero code changes | Done |
| Gateway | Thread-safe message routing between services | Done |
| Networking | Ethernet (QEMU) and WiFi (real hardware); HTTP client for API calls | Done |
| IM Integrations | Feishu (Lark) via WebSocket long connection; planned: DingTalk, QQ, Telegram | In Progress |
| Web Flash & Serial | Browser-based firmware flash (esptool-js) and serial terminal with ANSI color rendering | Done |
| Multi-Model API | Support mainstream LLM APIs: Claude, GPT, Gemini, DeepSeek, GLM, MiniMax, Grok, Moonshot, Baichuan, Qwen, Doubao, Llama (Ollama) | Planned |
| Web Config Portal | Lightweight built-in web page for configuring API keys, selecting models, and tuning parameters at runtime | Planned |
| Claw Skill Provider | Serve as a skill for other Claws, giving them the ability to sense and control the physical world | Planned |
+--------------------------------------------------------------+
| rt-claw Application |
| gateway | net | swarm | ai_engine | shell | sched | im |
+--------------------------------------------------------------+
| skills (AI Skills) |
| (one skill composes multiple tools) |
+--------------------------------------------------------------+
| tools (Tool Use) |
| gpio | system | lcd | audio | http | scheduler | memory |
+--------------------------------------------------------------+
| drivers (Hardware BSP) |
| WiFi | ES8311 | SSD1306 | serial | LCD framebuffer |
+--------------------------------------------------------------+
| osal/claw_os.h (OSAL API) |
+-------------------+------------------------------------------+
| FreeRTOS (IDF) | RT-Thread |
+-------------------+------------------------------------------+
| ESP32-C3 / S3 | QEMU vexpress-a9 |
+-------------------+------------------------------------------+
| Platform | Target | RTOS | Build | Status |
|---|---|---|---|---|
| ESP32-C3 | QEMU, xiaozhi-xmini, generic devkit | ESP-IDF + FreeRTOS | Meson + CMake | Verified |
| ESP32-S3 | QEMU, real hardware | ESP-IDF + FreeRTOS | Meson + CMake | Verified |
| vexpress-a9 | QEMU | RT-Thread | Meson + SCons | Verified |
# 1. Install ESP-IDF + QEMU (one-line setup)
./scripts/setup-esp-env.sh
# 2. Set your API key
export RTCLAW_AI_API_KEY='<your-api-key>'
# 3. Build and run on QEMU
make build-esp32c3-qemu
make run-esp32c3-qemuNo hardware? No problem. Try the CNB Cloud IDE — all toolchains pre-installed, build and run in your browser.
For real hardware (ESP32-S3/C3), WiFi setup, Feishu bot, and more — see the Getting Started Guide.
Join the GTOC (Gevico Open-Source Community) channels:
- Discord: RT-Claw
- QQ Group: Join
- Telegram: GTOC Channel
- Bilibili: Zevorn
- WeChat: GTOC
Inspired by OpenClaw, Nanobot, and MimiClaw.
MIT