A lightweight self-hosted AI agent written in Go that runs on your server and works through Telegram, using Anthropic, OpenAI or OpenRouter models.
zeroclaw-labs/zeroclaw
A Rust-based personal AI agent runtime you run on your own machine, connecting LLM providers to chat channels, tools and hardware.
- Stars
- 33,024
- Forks
- 4,985
- Open issues
- 1,035
- Last commit
- 3 hours ago
What is it?
ZeroClaw is a single Rust binary that runs an AI agent on your own machine. It connects to LLM providers such as Anthropic, OpenAI and Ollama, receives messages from channels including Discord, Telegram, Matrix, email, webhooks and a CLI, and takes actions through tools like shell, browser, HTTP and hardware GPIO/I2C/SPI access.
The agent defaults to a supervised autonomy mode in which medium-risk operations need approval and high-risk ones are blocked. It supports workspace boundaries, command policies, OS-level sandboxes (Landlock, Bubblewrap, Seatbelt, Docker), and a web dashboard with chat, memory browsing, config editing and cron management. Event-triggered procedures can be built with its SOP engine, and it can be used from editors over the Agent Client Protocol.
Configuration lives in a single TOML file, and the project offers a guided quickstart plus service installation for systemd, launchctl and Windows Services.
Key features
- 01Connects to Anthropic, OpenAI, Ollama and other OpenAI-compatible model providers with fallback chains.
- 02Receives and replies across Discord, Telegram, Matrix, email, webhooks and CLI through one agent loop.
- 03Defaults to supervised mode, requiring approval for medium-risk actions and blocking high-risk ones.
- 04Runs tools in OS-level sandboxes such as Landlock, Bubblewrap, Seatbelt and Docker.
- 05Provides an HTTP/WebSocket gateway and web dashboard for chat, memory, config and cron management.
- 06Supports event-triggered SOPs from MQTT, webhooks, cron and peripherals.
Use cases
- Running a personal assistant on your own machine that answers you over Telegram or Discord.
- Controlling Raspberry Pi or microcontroller hardware through natural-language commands.
- Connecting a local Ollama model to chat channels without sending data to a hosted service.
- Integrating an agent into an editor via the Agent Client Protocol.
Quick start
- 1
Run the official install script from the README
- 2
Run the guided setup to pick a provider and write a config
"${CARGO_HOME:-$HOME/.cargo}/bin/zeroclaw" quickstart - 3
Start an interactive chat with your configured agent
zeroclaw agent -a <alias>
Things to know
- You need an LLM provider API key or a local model server, such as Ollama, to set up an agent.
- Configuration is a multi-section TOML file, so initial setup takes some effort.
Topics
FAQ
- What is zeroclaw?
- A Rust-based personal AI agent runtime you run on your own machine, connecting LLM providers to chat channels, tools and hardware. ZeroClaw is a single Rust binary that runs an AI agent on your own machine. It connects to LLM providers such as Anthropic, OpenAI and Ollama, receives messages from channels including Discord, Telegram, Matrix, email, webhooks and a CLI, and takes actions through tools like shell, browser, HTTP and hardware GPIO/I2C/SPI access.
- Is zeroclaw free to use? What license does it have?
- zeroclaw is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is zeroclaw still maintained?
- Actively maintained. The last commit to zeroclaw was on Oct 22, 2026.
- What can I use zeroclaw for?
- Running a personal assistant on your own machine that answers you over Telegram or Discord. Controlling Raspberry Pi or microcontroller hardware through natural-language commands. Connecting a local Ollama model to chat channels without sending data to a hosted service. Integrating an agent into an editor via the Agent Client Protocol.
- How do I install zeroclaw?
- Run the official install script from the README Run the guided setup to pick a provider and write a config: "${CARGO_HOME:-$HOME/.cargo}/bin/zeroclaw" quickstart Start an interactive chat with your configured agent: zeroclaw agent -a <alias>
- Where can I get zeroclaw?
- The source code and installation instructions are on GitHub: https://github.com/zeroclaw-labs/zeroclaw
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An automation platform for building agentic AI workflows in YAML that run across conversational channels, with tools, MCP, memory and RAG.
Self-hostable personal AI assistant as a single Go binary, combining LLM chat, RAG knowledge bases, and agent loops for browsing, shell, and MCP tools.
A self-hosted personal AI assistant that runs on your machine or in the cloud and connects to chat apps like DingTalk, Lark, Discord and Telegram.
Open-source personal AI assistant and Agent Harness that plans tasks, runs tools and skills, and keeps memory across web and IM channels.
Open-source multi-agent AI assistant that runs tasks across web, desktop, files and chat apps like Telegram, Feishu and WeCom.