nullclaw
Fastest, smallest, and fully autonomous AI assistant infrastructure written in Zig
NullClaw is an ultra-lightweight, fully autonomous AI agent infrastructure written in Zig that compiles to a 678 KB static binary with sub-2ms startup and ~1 MB memory footprint. It provides a complete agent runtime with 50+ AI providers (OpenRouter, Anthropic, OpenAI, Gemini, Ollama, etc.), 19 messaging channels (Telegram, Discord, Signal, Slack, Nostr, IRC, WhatsApp, Matrix, and more), 35+ built-in tools, pluggable memory backends, MCP server integration, and hardware peripheral support. Every subsystem uses vtable interfaces for hot-swapping implementations without code changes, enabling developers to run production AI agents on $5 ARM boards or edge devices with minimal resource overhead.
Key Features
Use Cases
- 01Deploy autonomous AI assistants on edge devices and low-power ARM boards with minimal resource consumption
- 02Run multi-channel AI bots across Telegram, Discord, Signal, Slack, and Nostr with a single agent instance
- 03Build custom AI agents with swappable providers, memory backends, and tools without modifying core code
- 04Orchestrate agent workflows with topic-specific routing, named agent profiles, and cron-scheduled tasks
- 05Integrate hardware peripherals (Raspberry Pi GPIO, Arduino, STM32) with AI agent control loops
- 06Host secure, sandboxed agents with encrypted secrets and workspace-scoped file access for production deployments
Related Agents
View morehermes-agent
The agent that grows with you
agency-agents
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
openinterpreter
A coding agent for open models like Kimi K3
cline
Autonomous coding agent as an SDK, IDE extension, or CLI assistant.
nullclaw — FAQ
What is NullClaw?+
NullClaw is a fully autonomous AI agent infrastructure written in Zig that compiles to an ultra-small 678 KB static binary. It provides a complete runtime for building and deploying AI agents with pluggable providers, channels, tools, memory, and security sandboxing, optimized for edge devices and low-resource environments.
How do I install NullClaw?+
Install via Homebrew with 'brew install nullclaw', or build from source using Zig 0.16.0 with 'git clone https://github.com/nullclaw/nullclaw.git && cd nullclaw && zig build -Doptimize=ReleaseSmall'. After installation, run 'nullclaw onboard --interactive' to configure providers and channels.
Which AI providers does NullClaw support?+
NullClaw supports 50+ providers including OpenRouter, Anthropic, OpenAI, Azure OpenAI, Gemini, Vertex AI, Ollama, Groq, Mistral, xAI, DeepSeek, Together, Fireworks, Perplexity, Cohere, Bedrock, and any OpenAI-compatible custom endpoint. Providers are configured in ~/.nullclaw/config.json under models.providers.
What channels and messaging platforms can NullClaw connect to?+
NullClaw supports 19 channels: CLI, Telegram, Signal, Discord, Slack, iMessage, Matrix, WhatsApp, Webhook, IRC, Lark/Feishu, OneBot, Line, DingTalk, Email, Nostr, QQ, MaixCam, and Mattermost. Each channel is configured under 'channels' in config.json with account-specific credentials and allowlists.
Is NullClaw free and open source?+
Yes, NullClaw is fully open source under the MIT license with no tokens, cryptocurrency, or paid tiers. The project has no financial instruments and is a pure open-source infrastructure project.
What are the prerequisites to run NullClaw?+
NullClaw requires only libc and optionally SQLite for the default memory backend. To build from source you need Zig 0.16.0 exactly. API keys are needed for your chosen AI provider (e.g., OpenRouter, Anthropic) and for messaging channels (Telegram bot token, Discord token, etc.).
How do I install nullclaw?+
Open the source repository on GitHub and follow its README. nullclaw is a agent — MCP Agents Market links you directly to the official repo.
Is nullclaw free?+
nullclaw is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.