zeroclaw-labs/zeroclaw

Fast, small, and fully autonomous AI personal assistant infrastructure, any OS, any platform — deploy anywhere, swap anything 🦀

What it solves

ZeroClaw is a self-hosted agent runtime that allows users to own their AI assistant, their data, and the machine it runs on. It eliminates reliance on centralized AI platforms by providing a single binary that can be configured to use various LLM providers and interact with the world through multiple communication channels and tools.

How it works

Written in Rust, ZeroClaw acts as a central hub that connects LLM providers (such as Anthropic, OpenAI, and Ollama) to various input/output channels (Discord, Telegram, Matrix, email, etc.) and actionable tools (shell, browser, HTTP, and hardware peripherals). It includes a security policy engine to manage autonomy levels, sandboxing (via Landlock, Bubblewrap, etc.), and cryptographic tool receipts for every action taken by the agent.

Who it’s for

It is designed for users who prioritize privacy, security, and local control over their AI agents, as well as developers who want to integrate AI into their local environment, hardware (like Raspberry Pi), or specific communication channels.

Highlights

  • Multi-channel support: Access a single agent across 30+ channels including Discord, Telegram, and a custom CLI.
  • Provider-agnostic: Pluggable model providers with fallback chains and routing to ensure reliability.
  • Security-first: Implements supervised autonomy, OS-level sandboxing, and tool receipts for safety.
  • Hardware-capable: Direct interaction with GPIO, I2C, SPI, and USB on platforms like Raspberry Pi and Arduino.
  • SOP engine: Event-triggered Standard Operating Procedures for automated tasks with approval gates.
  • IDE Integration: Supports the Agent Client Protocol (ACP) for editor integration via JSON-RPC.

Related

  • Project
    qhkm/zeptoclaw

    A fast, secure, and local-first personal AI assistant infrastructure written in Rust, providing sandboxed autonomy and multi-provider LLM support in a single 6MB binary.

  • Project
    thClaws/thClaws

    A native-Rust AI agent workspace that allows users to run multiple isolated agents for coding and automation locally on their own machine.

  • Project
    microclaw/microclaw

    MicroClaw is a self‑hosted Rust platform for AI agents that can run as a server, a native desktop “Work” app, or be embedded via an SDK. It offers multi‑step tool use, durable sessions, SQLite‑based memory, many chat channel adapters, and built‑in safety guardrails, all share‑able across its three product surfaces.

  • Project
    ClawRunr/JavaClaw

    A Java-based personal AI assistant that runs locally to manage tasks, execute shell commands, and communicate across multiple channels like Discord and Telegram.

  • Project
    fastclaw-ai/fastclaw

    A lightweight AI agent runtime written in Go that allows users to create, manage, and deploy multi-agent systems with sandboxed tool execution and IM channel integrations.