Mid-weight Python AI Agent with rich ecosystem support. Runs on $50 SBCs like Raspberry Pi.
Part of the Clawland ecosystem.
NanoClaw bridges the gap between the ultra-lightweight PicClaw and the full-featured MoltClaw. Built in Python, it leverages the massive Python ecosystem for ML, computer vision, data processing, and automation — all on affordable single-board computers.
- Python Ecosystem — Full access to NumPy, OpenCV, TensorFlow Lite, scikit-learn, and more
- Local ML Inference — Run small models directly on edge hardware
- Rich I/O — GPIO, I2C, SPI, Serial, Camera, Microphone support
- Agent Capabilities — Tool use, memory, multi-step reasoning
- Cloud Sync — Report to MoltClaw, receive orchestration commands
| Component | Minimum | Recommended |
|---|---|---|
| RAM | 100MB | 512MB+ |
| Storage | 200MB | 1GB+ |
| Hardware | Raspberry Pi Zero 2W | Raspberry Pi 4/5 |
| Cost | ~$15 | ~$50 |
- Smart Camera — Person/object detection with Pi Camera + TFLite
- Voice Assistant — Local wake-word + cloud LLM hybrid
- Data Collector — Aggregate sensor data from multiple MicroClaw nodes
- Lab Monitor — Temperature, humidity, air quality with ML anomaly detection
🚧 Pre-Alpha — Architecture design phase. Looking for contributors!
See the Clawland Contributing Guide.
Core contributors share 20% of product revenue. Read the Contributor Revenue Share terms.
Apache License 2.0 — see LICENSE for details.