Zero setupOne local engine, nothing to configure
Neatype runs a single, fixed model on your Mac and sets it up for you — no model picker, no tuning, no decisions to make.
Set up automatically
On first launch Neatype downloads the local runtime (~133 MB) and the model (~2.5 GB) and starts them. You install nothing.
Stays on your Mac
The model and all inference run on 127.0.0.1 — and if you already run Ollama, Neatype reuses it instead of downloading its own.
Tuned for speed
A compact model with its “thinking” step switched off, so a correction comes back in well under a second.
The model lives in Application Support, and the local server shuts down the moment you quit Neatype.