Repolode
Submit
sipeed

sipeed/picoclaw

A lightweight personal AI assistant written in Go that runs on low-cost hardware and connects to many LLM providers and chat apps.

View on GitHubWebsite
Stars
30,082
Forks
4,465
Open issues
39
Last commit
5 days ago

What is it?

PicoClaw is a personal AI assistant written in Go as a single binary. It targets low-resource Linux devices, with a stated memory footprint under 10MB for the core and boot times under a second on slow single-core processors. It is built for x86_64, ARM64, MIPS, RISC-V and LoongArch, and the project lists tested boards from very cheap RISC-V boards to Raspberry Pi and Android phones.

The assistant connects to 30+ LLM providers through a model_list configuration, including OpenAI, Anthropic, Gemini, Ollama and vLLM for local models. It can be reached through 19+ messaging platforms such as Telegram, Discord, Slack, Matrix, Feishu and WeCom. It also supports MCP servers for external tools, web search through several engines, scheduled reminders via cron, and skills loaded from SKILL.md files.

Users can set it up through a browser-based WebUI launcher, Docker Compose, an Android APK, or the command-line interface, which offers one-shot and interactive chat as well as gateway mode for chat apps.

Key features

  • 01Single Go binary that runs across x86, ARM, MIPS, RISC-V and LoongArch.
  • 02Connects to 30+ LLM providers, including local Ollama and vLLM.
  • 03Supports 19+ chat platforms such as Telegram, Discord, Slack and Feishu.
  • 04Native MCP support for connecting external tool servers.
  • 05Scheduled reminders and recurring tasks through a cron tool.
  • 06Browser-based WebUI launcher for configuration and chat.

Use cases

  • Running a personal AI assistant on a cheap single-board computer or old Android phone.
  • Home automation or surveillance assistant on embedded camera or RISC-V boards.
  • Chatting with an LLM agent from Telegram, Slack or other messaging apps.
  • Automating recurring reminders and scheduled tasks from a self-hosted assistant.

Quick start

  1. 1

    Download the app from picoclaw.io (the site detects your platform)

  2. 2

    Start the WebUI Launcher from the command line

    picoclaw-launcher
  3. 3

    Open http://localhost:18800 in your browser, then configure a provider and channel

Things to know

  • The project is in early rapid development and may have unresolved security issues.
  • The README advises against production use before v1.0.
  • Recent builds may use 10-20MB of RAM instead of the advertised under 10MB.

Topics

#ai-assistant#golang#llm#mcp#chatbot#embedded#self-hosted#edge-computing

FAQ

What is picoclaw?
A lightweight personal AI assistant written in Go that runs on low-cost hardware and connects to many LLM providers and chat apps. PicoClaw is a personal AI assistant written in Go as a single binary. It targets low-resource Linux devices, with a stated memory footprint under 10MB for the core and boot times under a second on slow single-core processors. It is built for x86_64, ARM64, MIPS, RISC-V and LoongArch, and the project lists tested boards from very cheap RISC-V boards to Raspberry Pi and Android phones.
Is picoclaw free to use? What license does it have?
picoclaw is open source under the MIT license. Check the license terms before using it in commercial products.
Is picoclaw still maintained?
Actively maintained. The last commit to picoclaw was on Oct 15, 2026.
What can I use picoclaw for?
Running a personal AI assistant on a cheap single-board computer or old Android phone. Home automation or surveillance assistant on embedded camera or RISC-V boards. Chatting with an LLM agent from Telegram, Slack or other messaging apps. Automating recurring reminders and scheduled tasks from a self-hosted assistant.
How do I install picoclaw?
Download the app from picoclaw.io (the site detects your platform) Start the WebUI Launcher from the command line: picoclaw-launcher Open http://localhost:18800 in your browser, then configure a provider and channel
Where can I get picoclaw?
The source code and installation instructions are on GitHub: https://github.com/sipeed/picoclaw

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects