A local LLM application that lets you download open source models and run them entirely on your own hardware.
- Lightweight and performant -- Rust backend with llama.cpp inference and Metal GPU acceleration. Tauri v2 keeps the app small.
- Private -- Fully offline after model download. No telemetry, no accounts, no data leaves your machine.
- Free and open source -- MIT licensed. No vendor lock-in, no proprietary dependencies.
- Architecture -- System components, data flows, and repository structure.
- Design Decisions -- Technology choices and the rationale behind them.