zeroclaw  by zeroclaw-labs

Autonomous AI assistant infrastructure for minimal deployments

Created 5 days ago

New!

10,057 stars

Top 5.0% on SourcePulse

GitHubView on GitHub
Project Summary

Fast, small, and fully autonomous AI assistant infrastructure, ZeroClaw addresses the need for highly efficient, deployable AI agents. It targets engineers and power users seeking to run AI assistants on minimal hardware with maximum flexibility, offering a zero-overhead, zero-compromise solution.

How It Works

Built entirely in Rust, ZeroClaw prioritizes performance and minimal resource consumption. Its core architecture is designed around traits, allowing any subsystem—including AI model providers, communication channels, memory backends, and tools—to be swapped out with zero code changes. This pluggable design, combined with a single, self-contained binary, ensures true portability across architectures (ARM, x86, RISC-V) and enables deployment on resource-constrained devices.

Quick Start & Requirements

Installation involves cloning the repository, building with cargo build --release --locked, and then installing with cargo install --path . --force --locked. Key commands include zeroclaw onboard for setup, zeroclaw agent for interaction, and zeroclaw gateway or zeroclaw daemon for running services. Prerequisites include the Rust toolchain and platform-specific build tools (Visual Studio Build Tools on Windows, build-essential on Linux, Xcode Command Line Tools on macOS). Docker is optional for sandboxed runtimes.

Highlighted Details

  • Ultra-Lightweight: Operates with a <5MB memory footprint and a ~3.4MB binary size.
  • Minimal Cost: Capable of running on $10 hardware, significantly cheaper than comparable solutions.
  • Lightning Fast: Achieves sub-10ms startup times, even on low-clocked hardware.
  • Extensive Swappability: Supports 22+ AI model providers (including OpenAI-compatible endpoints), various communication channels (CLI, Telegram, Discord, Slack), and multiple memory backends (SQLite with hybrid search, Lucid, Markdown).
  • Security Focused: Implements gateway pairing, strict sandboxing, filesystem scoping, and explicit allowlists by default.

Maintenance & Community

The project is maintained by zeroclaw-labs. While the README expresses gratitude to institutions like Harvard and MIT, it does not detail specific maintainers, sponsorships, or direct community channels like Discord or Slack.

Licensing & Compatibility

ZeroClaw is licensed under the Apache 2.0 license, which permits commercial use and integration into closed-source projects.

Limitations & Caveats

WebAssembly (WASM) runtimes are planned but currently unsupported. The browser_open and browser tools, particularly the computer_use backend, require external WebDriver servers or sidecar applications to function. WhatsApp integration necessitates a secure HTTPS connection, typically via a tunnel.

Health Check
Last Commit

14 hours ago

Responsiveness

Inactive

Pull Requests (30d)
558
Issues (30d)
204
Star History
11,199 stars in the last 5 days

Explore Similar Projects

Starred by Peter Norvig Peter Norvig(Author of "Artificial Intelligence: A Modern Approach"; Research Director at Google), Zhen Lu Zhen Lu(Cofounder of Runpod), and
1 more.

agents-towards-production by NirDiamant

0.7%
17k
Production-ready GenAI agent tutorials
Created 8 months ago
Updated 1 day ago
Feedback? Help us improve.