PicoClaw
A tiny Go-based personal AI assistant that runs on cheap hardware with minimal memory.
- Stars
- 30.1k
- Forks
- 4.4k
- Licence
- MIT
- Language
- Go
- Last update
- 6 days ago
- Type
- AI model / agent
What it is
PicoClaw is an open-source personal AI assistant written in Go, designed to use very little memory and run on low-cost boards. The README says it is inspired by NanoBot and aims to automate everyday tasks through an AI agent. It supports features like MCP server connections, image input, and rule-based model routing.
- Run a personal assistant on a cheap single-board computer
- Connect MCP servers to extend what the agent can do
- Send images and files to the agent for multimodal LLM processing
- Set up model routing so simple queries use lighter models
Great for
- Hobbyists who want an AI assistant on a low-cost Linux board
- Developers experimenting with lightweight AI agents in Go
- People who want to connect MCP servers to an assistant
- Users with limited hardware such as RISC-V or ARM single-board computers
Not ideal for
- Production deployments, because the README says not to deploy before v1.0 and that security issues may remain
- Users who need guaranteed low memory use, since the README notes recent builds may use 10-20MB RAM
- Non-technical users looking for a polished consumer app, as setup details are mostly developer-oriented
How to get it
- 1
Download the app
Clone the repository
git clone https://github.com/sipeed/picoclaw - 2
Run it
- Download the release archive for your platform from the GitHub releases page
- Extract the archive and locate the picoclaw binary
- Consult the official docs at docs.picoclaw.io for configuration and first run
- 3
Get the most out of it
- Use the Web UI launcher and Docker Compose support mentioned in the release notes
- Add MCP servers via the built-in MCP CLI commands (show, add, list, remove, test, edit)
- Enable configurable web search as described in the v0.2.7 and v0.2.9 notes
- Check the Hardware Compatibility List to confirm your board is tested
Downloads
Strengths
- Written in Go and distributed as a single binary across many architectures, per the README
- Active development with frequent releases, including v0.3.1 listed as latest
- MIT licensed
Trade-offs
- The README warns it is in early rapid development and may have unresolved security issues
- Memory footprint claims (under 10MB) are caveated; recent builds may use 10-20MB
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| picoclaw (this) | 30.1k | AI model / agent | intermediate | 6 days ago | Hobbyists who want an AI assistant on a low-cost Linux board |
| nanobot | 49.8k | App | intermediate | 4 days ago | People who want a self-hosted personal assistant they can inspect and customize |
| openclaw | 394k | App | intermediate | 4 days ago | People who want an AI assistant that runs on their own machine rather than a hosted service |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.