Python. SQLite. MLX. A local-first AI framework that gives you complete sovereignty over your data and privacy. Run powerful AI models on your own hardware.
Everything you need to build autonomous AI, nothing you don't.
Pure Python with async support. No heavy frameworks. FastAPI-compatible API layer for integrations.
Embedded database for zero-config persistence. Vector extensions for semantic memory and retrieval.
Hardware-optimized inference. Run 7B-70B parameter models locally with GPU acceleration on your machine.
Control Whity from anywhere via Telegram. Send commands, receive updates, and manage your AI assistant remotely with seamless bot integration.
A recursive self-improvement loop at the core of everything.
Voice, text, code, images
MLX inference engine
SQLite + vector store
Self-evaluation cycle
Tools, APIs, output
Toggle between cloud-powered and fully local at any time.
Full power. Connect to external APIs, use cloud models alongside local ones, and access web capabilities.
Maximum privacy. Everything runs on your hardware. Zero network requests. Your data never leaves your machine.