Ultra-efficient personal AI assistant powered by Crystal
Autobot is a personal AI assistant framework written in Crystal — inspired by OpenClaw, rebuilt from scratch with security and efficiency as first-class concerns.
- 2MB binary, ~5MB RAM, <20ms startup, zero runtime dependencies
- Multi-provider LLM — Anthropic, OpenAI, DeepSeek, Groq, Gemini, OpenRouter, AWS Bedrock, DuckAI, vLLM
- Chat channels — Telegram, Slack, WhatsApp, Zulip with allowlists and custom commands
- Kernel-enforced sandboxing — Docker/bubblewrap OS-level isolation with custom
Dockerfile.sandbox - Vision and voice — image analysis and Whisper transcription
- Builtin plugins — SQLite, GitHub, Weather; opt-out via config
- Persistent memory, cron scheduling, MCP servers, markdown skills
- Run dozens of isolated bots on a single machine
brew tap crystal-autobot/tap && brew install autobot
autobot new optimus && cd optimus
autobot doctor
autobot gateway| Repository | Description |
|---|---|
| ark | Slack gateway for AWS Bedrock Agents via Socket Mode |
| autobot | Core framework — agent runtime, providers, channels, plugins, tools |
| autobot-wasm | WebAssembly plugin — write tools in any language, compile to .wasm |
| blueprints | Ready-to-use templates — Optimus, Bumblebee, Blaster, Red Alert |
| homebrew-tap | Homebrew formula for macOS/Linux installation |
