ZeroClaw

ZeroClaw is a personal AI agent for people who want an assistant running on their own computer or server. It can use a shell and browser, respond through messaging apps and perform scheduled tasks.

• Model choice Connect Ollama, LM Studio, llama.cpp or vLLM for local inference, or use Anthropic, OpenAI and OpenAI-compatible endpoints, with fallback routing between providers.
• Messaging and automation Reach the agent through Telegram, Discord, Signal, Matrix or email, and trigger work through schedules, webhooks and channel events.
• Action controls Default supervised mode requires approval for medium-risk actions and blocks high-risk ones; command allowlists, workspace boundaries and operating system sandboxes constrain access.
• Dashboard and tools Use the web dashboard for chat, memory browsing and task management, and extend the agent with custom MCP servers or hardware tools for Raspberry Pi, Arduino and ESP32.

The agent is a single Rust binary for Linux, macOS and Windows, with Docker support and a runtime small enough for Raspberry Pi hardware. It's free and open source under Apache 2.0.

With a local model, inference stays on your hardware. Hosted models receive prompts through your own provider key, without an intervening ZeroClaw cloud service. Messaging channels still involve the services you connect. Optional cryptographic receipts help verify successful tool calls.