macOS · local · open source

FLOWY

Hold a hotkey. Speak. Release.
Your words appear — in any window, on your machine.

privacy

No cloud.
Not ever.

Every byte of audio stays on your machine. No API key, no account, no subscription.

engine

Apple Neural.
Built in.

cost

Free.
Always.

MIT license, open source. No tier that removes features. No catch.

capabilities

What it does

Push-to-talk Hold any global hotkey, speak from any app — email, code editor, browser, terminal. Release to transcribe. Text injects directly into the focused window.
Native recognition Apple SFSpeechRecognizer — built into macOS. No warm-up time, no model files. on-device only
Custom dictionary Word substitutions applied after every transcription. Teach Flowy your names, abbreviations, and jargon. Test them live in the settings panel.
Local AI polish Optional pass through a local Ollama model to fix punctuation and grammar. Still entirely offline — no cloud round-trips ever. ollama required
Lightweight Under 15 MB. Zero CPU when idle. Sits silently in your menu bar until you need it. Starts at login.

getting started

Install

Download the .dmg, drag to Applications, open it. Grant microphone and speech recognition access when macOS prompts you.
Set your key

Pick any global shortcut in Settings. The default works immediately.
⌘ Shift Space
Speak anywhere

Focus any text field. Hold your hotkey. Speak. Release. Words appear — every time.

macOS 11 Big Sur or later · Apple Silicon & Intel

Download .dmg MIT License · open source · no telemetry · no subscription