# ZeroClaw
ZeroClaw is an open-source AI assistant framework built entirely in [[Rust]], designed for minimal resource consumption and maximum flexibility. Created by ZeroClaw Labs, it focuses on "zero overhead" deployment, running on hardware as cheap as a $10 [[Raspberry Pi]] while consuming less than 5MB of RAM.
> Deploy anywhere, swap anything.
## How It Relates to OpenClaw
ZeroClaw is a Rust-based alternative to [[OpenClaw]]. While OpenClaw is built on Node.js and focuses on messaging-app integrations and orchestration, ZeroClaw prioritizes extreme efficiency, tiny binaries, and bare-metal deployment. Both projects share the "claw" lineage and support the [[OpenClaw]] identity format, but they target different ends of the spectrum: OpenClaw for feature-rich orchestration, ZeroClaw for lean, embeddable infrastructure.
## Key Features
- **Lean runtime**: Less than 5MB RAM on release builds, near-instant cold starts from a single binary
- **Provider agnostic**: 29+ built-in AI model providers plus custom endpoints
- **Multi-channel**: Telegram, Discord, Slack, CLI, email, and 10+ other platforms
- **Trait-driven architecture**: Every subsystem (AI providers, communication channels, memory backends) can be swapped without code changes
- **Security-first**: Pairing, sandboxing, explicit allowlists, and workspace scoping
- **Custom memory engine**: SQLite-based vector search with FTS5 keyword indexing, no external dependencies
## Hardware Requirements
- Minimum for compilation: 2GB RAM + 6GB disk
- Runtime: Can operate on $10 hardware (e.g., Raspberry Pi)
- Pre-built binaries available for x86_64, ARM64, and ARMv7
## Installation
### Homebrew
```bash
brew install zeroclaw
```
### One-click Bootstrap
```bash
curl -fsSL https://zeroclawlabs.ai/install.sh | bash
```
### From Source
```bash
git clone https://github.com/zeroclaw-labs/zeroclaw.git
cd zeroclaw
cargo build --release
```
## Architecture
ZeroClaw uses a modular trait-based architecture in Rust. Key subsystems include:
- **Memory**: Custom SQLite-based vector search with FTS5 keyword indexing
- **Runtime**: Native execution and Docker sandboxing
- **Tools**: Shell, file operations, git, HTTP requests, scheduling, and hardware integrations
- **Identity**: [[OpenClaw]] markdown and AIEOS JSON format support
## References
- GitHub: https://github.com/zeroclaw-labs/zeroclaw
- Website: https://zeroclawlabs.ai/
## Related
- [[OpenClaw]]
- [[AI Agents]]
- [[Rust]]
- [[Claude Code]]