ZeroClaw — Lightweight Personal AI Agent

Description

ZeroClaw is a personal AI assistant designed to run on your own machine, providing a secure and private AI experience. It operates as a single Rust binary, eliminating the need for additional runtime dependencies such as Node, JVM, or Python environments. This lightweight architecture allows ZeroClaw to start in milliseconds and use less memory than a browser tab, making it suitable for devices as small as a Raspberry Pi.

ZeroClaw supports over 30 communication channels, including popular platforms like Telegram, Discord, WhatsApp, Slack, and email. It connects to more than 70 LLM providers, offering flexibility in choosing between local models like Ollama and hosted providers such as OpenAI and Anthropic. The agent ensures that your data remains on your machine, with only the prompts sent to your chosen provider when using hosted services.

Security is a core feature of ZeroClaw, with supervised autonomy as the default mode. Medium-risk actions require user approval, while high-risk actions are blocked. The agent is contained within OS-level sandboxes like Landlock, Bubblewrap, Seatbelt, or Docker, and employs command allowlists and workspace scoping to control its operations. Users can opt-in for cryptographic tool receipts, providing verifiable proof of tool calls.

ZeroClaw is open source and dual-licensed under MIT or Apache-2.0, with no subscription fees. Users only incur costs from their chosen LLM providers, or none if using a local model. The agent is designed for various use cases, from personal assistants to unattended automation, offering a customizable and private AI solution.

ZeroClaw's Core Features

• Single Rust binary

• Supports 30+ channels

• Connects to 70+ LLM providers

• Runs locally on your machine

• Supervised autonomy by default

• OS sandboxing for security

• Open source, MIT or Apache-2.0 licensed

• Opt-in cryptographic tool receipts

• Persistent memory

• No runtime dependencies

How to use ZeroClaw?

1. 1

   Install: Download and install the Rust binary

2. 2

   Configure: Set up your keys or local models

3. 3

   Connect: Link to your preferred channels and providers

4. 4

   Use: Run the agent for personal AI tasks

5. 5

   Secure: Enable supervised autonomy and sandboxing

ZeroClaw's Use Cases

• Personal AI Assistant
• Code Generation
• Data Analysis
• Document Processing
• Task Automation

FAQ from ZeroClaw

• What is ZeroClaw?

  ZeroClaw is a personal AI assistant that runs as a single Rust binary on your machine, supporting 70+ LLM providers and 30+ channels.

• Is ZeroClaw free?

  Yes, ZeroClaw is open source and free to use. You only pay for your LLM provider costs, if any.

• How does ZeroClaw ensure data privacy?

  ZeroClaw runs locally on your machine, ensuring that data does not leave your device unless you use a hosted provider.

• What security features does ZeroClaw offer?

  ZeroClaw includes supervised autonomy, OS sandboxing, command allowlists, and optional cryptographic tool receipts.

• What platforms does ZeroClaw support?

  ZeroClaw runs on ARM and x86 architectures, including devices like Raspberry Pi.

• Can ZeroClaw connect to multiple channels?

  Yes, ZeroClaw supports over 30 channels, including Telegram, Discord, and email.

• What are the licensing terms for ZeroClaw?

  ZeroClaw is dual-licensed under MIT or Apache-2.0, allowing for flexible use and modification.

• How lightweight is ZeroClaw?

  ZeroClaw is a single native binary that starts in milliseconds and uses less memory than a browser tab.