A privacy-first desktop agent that turns natural-language requests into verified on-screen actions through a local perception and planning loop.
- Built a perceive, understand, plan, act, and verify loop with LLaVA, Mistral, PaddleOCR, and PyAutoGUI.
- Engineered a Rust and Tauri runtime alongside a FastAPI and WebSocket backend.
- Python
- Rust
- Tauri
- Next.js
- FastAPI
- PyAutoGUI
- LLaVA
- Mistral
- PaddleOCR