ZeroClaw Puts a Full AI Agent on Your Own Machine for Free
ZeroClaw ships a single Rust binary that runs an autonomous personal AI agent across 30+ channels, 20+ LLM providers, and even $10 microcontrollers.
PRO- ZeroClaw is a single Rust binary personal AI agent runtime, MIT/Apache-2.0 licensed.
- Connects to ~20 LLM providers including Anthropic, OpenAI and local Ollama, with fallback routing.
- Reaches users through 30+ channels: Discord, Telegram, Matrix, email, voice, webhooks, CLI.
- Supervised autonomy by default with OS-level sandboxes (Landlock, Bubblewrap, Seatbelt, Docker) and cryptographic tool receipts.
- Runs on cheap hardware: kernel-only preset is ~6.6 MB, supports Raspberry Pi, STM32, Arduino and ESP32 peripherals.
- SOP engine triggers procedures from MQTT, webhook, cron, or peripheral events with approval gates.
A new open-source project called ZeroClaw argues that a personal AI assistant does not need to live in someone else's cloud. It is a single Rust binary you install on your own machine, wire up to whatever LLM you want, and expose through whatever messaging channels you already use.
The repo has climbed past 30,000 stars and is dual-licensed under MIT and Apache 2.0, with an install script that lets you pick a prebuilt binary or a source build in under a minute.
One binary, many mouths to feed
ZeroClaw is an agent runtime: a single Rust binary you configure and run. The interesting part is its surface area. It talks to LLM providers like Anthropic, OpenAI, Ollama and roughly 20 others, reaches the world through 30+ channels such as Discord, Telegram, Matrix, email, voice, webhooks, and your own CLI, and acts through tools including shell, browser, HTTP, hardware, and custom MCP servers.
Everything runs locally with your own API keys inside a workspace you control, which is the core pitch: you own the agent, you own the data, you own the machine it runs on.
Security that is not just a checkbox
Autonomous agents that can run shell commands are terrifying by default, and the project takes that seriously. The default autonomy level is supervised, meaning medium-risk operations require approval and high-risk ones are blocked outright. On top of that sit workspace boundaries, a command policy, OS-level sandboxes using Landlock, Bubblewrap, Seatbelt or Docker depending on your platform, and cryptographic tool receipts on every action.
This story is for Pro members
You've reached the end of the free preview. Upgrade to AlphaSignal Pro to read the full article - and everything else behind the paywall.
Comments
