A personal AI assistant built in Rust, ZeptoClaw addresses the tradeoffs found in previous AI agent frameworks by offering a unified, ultra-lightweight solution. It targets engineers and power users seeking a performant, secure, and resource-efficient AI agent, delivering a comprehensive feature set within a minimal footprint.

How It Works

Engineered in Rust for memory safety and async performance, ZeptoClaw utilizes a robust, multi-layered security architecture that operates by default. This includes container isolation (Docker/Apple Container), prompt injection detection, secret scanning, policy enforcement, and input validation. It supports multi-provider LLMs (Claude, OpenAI) with automatic failover and integrates numerous tools, all within a single, small binary.

Quick Start & Requirements

Installation is streamlined via a one-liner curl script, Homebrew (brew install qhkm/tap/zeptoclaw), Docker (docker pull ghcr.io/qhkm/zeptoclaw:latest), or direct source compilation (cargo install). Container isolation requires Docker or Apple Container. The project boasts a minimal 4MB binary size and 6MB RAM usage. Official documentation is available at zeptoclaw.com/docs.

Highlighted Details

• Multi-provider LLM support (Claude, OpenAI) with SSE streaming and auto-failover.
• Extensive toolset including shell, filesystem, web, memory, cron, and integrations like WhatsApp and Google Sheets.
• Agent swarms for delegating tasks to specialized sub-agents.
• Comprehensive, default-enabled security features: container isolation, prompt injection detection, secret scanning, policy engine, input validation, shell blocklist, SSRF prevention, and tool approval gate.
• Efficient multi-tenant deployment on a single VPS, with isolated workspaces consuming ~6MB RAM each.
• A migration utility for transitioning from OpenClaw configurations.

Maintenance & Community

Developed by Aisar Labs. Specific community channels or detailed contributor information were not detailed in the provided README snippet.

Licensing & Compatibility

Released under the Apache 2.0 license, which is permissive for commercial use and integration into closed-source projects.

Limitations & Caveats

While ZeptoClaw aims to eliminate tradeoffs, the effectiveness of its extensive security layers against novel or sophisticated threats would require independent validation. Specific limitations regarding unsupported platforms or advanced feature gaps were not explicitly stated in the provided text.