Skip to content

About

Lightweight local voice dictation for macOS using Whisper

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

LocalFlow

LocalFlow

Voice-to-text that runs entirely on your Mac

Features • Installation • Usage • Configuration • License

Platform Chip License


LocalFlow is a privacy-first voice dictation app for macOS. Speech recognition runs 100% locally using OpenAI's Whisper model, no internet connection required, no data ever leaves your device.

LocalFlow recording overlay

Features

  • Double-tap to dictate: Double-tap Option key, hold and speak, release to transcribe
  • Works in any app: Text is inserted wherever your cursor is
  • Auto-updates: Built-in update system keeps you current
  • Visual feedback: Minimal floating pill with live waveform and timer; hover it to see the live text
  • Self-corrections: Say "John, I mean Sarah" and only "Sarah" is kept
  • Battery friendly: Live transcription only runs while you hover the pill
  • Clipboard friendly: Your clipboard is restored and dictations stay out of clipboard managers
  • Punctuation mode: Automatically add punctuation
  • Clipboard mode: Copy to clipboard without auto-pasting
  • History: Access your recent transcriptions
  • Custom hotkey: Configure your preferred trigger key
  • Sound feedback: Audio cues for recording start/stop
  • Auto-launch: Start LocalFlow when you log in
  • Fast: Sub-second transcription on Apple Silicon
  • Private: No cloud, no telemetry, no data collection

Requirements

  • macOS 14.0+ (Sonoma)
  • Apple Silicon (M1/M2/M3/M4)

Installation

Download (Recommended)

  1. Download the latest DMG from Releases

  2. Open the DMG and drag LocalFlow to Applications

  3. First launch: Right-click LocalFlow.app and select "Open"

    macOS will warn that the app is from an unidentified developer. Click "Open" to proceed. This only happens once.

  4. Grant permissions when prompted:

    • Accessibility: Required for hotkey detection and text insertion
    • Microphone: Required for voice recording
  5. Download a model: Go to Settings > Model and download your preferred model

That's it! Double-tap Option to start dictating.

Bypassing Gatekeeper

Since LocalFlow isn't notarized with Apple, macOS shows a warning on first launch. Two ways to handle this:

Option A: Right-click to Open

  1. Right-click (or Control-click) on LocalFlow.app
  2. Select "Open" from the menu
  3. Click "Open" in the dialog

Option B: System Settings

  1. Try to open LocalFlow normally (it will be blocked)
  2. Go to System Settings > Privacy & Security
  3. Scroll down to find "LocalFlow was blocked"
  4. Click "Open Anyway"

Build from Source

For developers who want to build from source, see docs/INSTALL.md.

Usage

Action How
Start recording Double-tap Option key, then hold
Stop & transcribe Release the key
Open menu Click waveform icon in menu bar
Settings Click menu bar icon → Settings
Check for updates Click menu bar icon → Check for Updates

Configuration

Access settings via the menu bar icon.

Setting Description
Model Choose Whisper model (tiny/base/small/medium)
Hotkey Configure trigger key (Option, Control, Fn)
Punctuation Auto-add punctuation
Clipboard mode Copy only, don't paste
Sound feedback Play sounds on record start/stop
Launch at login Start automatically

Models

Model Size Speed Accuracy
tiny 75MB ★★★★★ ★★☆☆☆
base 142MB ★★★★☆ ★★★☆☆
small 466MB ★★★☆☆ ★★★★☆
medium 1.5GB ★★☆☆☆ ★★★★★

Recommendation: Start with small for the best balance of speed and accuracy.

Documentation

Privacy

LocalFlow is designed with privacy as a core principle:

  • All speech recognition happens on-device using whisper.cpp
  • No internet connection required for transcription
  • No data is ever sent to external servers
  • No analytics or telemetry
  • Update checks only fetch a small XML file from GitHub

See our full Privacy Policy.

License

MIT License: see LICENSE for details.


Made with Swift for macOS

About

Lightweight local voice dictation for macOS using Whisper

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages