# SayIt

Open-source voice typing for Windows: press a shortcut, speak, and AI-cleaned-up text is inserted at the cursor, using local models, the user's own API keys, or a self-hosted backend.

SayIt transcribes speech, cleans it up and types the result wherever the cursor is, working in editors, chat apps, browsers and other Windows software without switching windows. Every cleanup prompt is editable, letting a user remove filler words, fix recognition errors, or keep a faithful transcript; an optional context-aware mode reads text around the cursor so new dictation matches its tone, and can turn selected text plus speech into an editing instruction such as translate, tighten or rewrite.

It offers three data-flow modes: fully local recognition on the user's PC with no data leaving the device when AI cleanup is off, a cloud-API mode that talks directly to the chosen speech-recognition and AI providers, and a server mode for teams that runs through a self-hosted or public trial backend. Local mode ships several GGUF speech-recognition models with GPU acceleration where available, and the app supports hotwords, per-app cleanup rules, and an English and Chinese interface.

Website: https://sayitapp.site