Skip to content

draenger/whisperio

Whisperio

Whisperio

Global dictation for Windows, macOS, and Linux — press a hotkey, speak, and your words are transcribed and auto-pasted into any app.

Download · Website · Sponsor


How it works

  1. Press a global hotkey (works from any application)
  2. Speak into your microphone — a minimal overlay shows recording status
  3. Press the hotkey again to stop
  4. Your speech is transcribed via cloud API and instantly pasted into the focused app

No browser tabs, no copy-paste — just talk and it types.

Features

  • Global hotkey — works system-wide, even when Whisperio is in the background
  • Auto-paste — transcription is placed directly into the focused text field
  • Dictate & Send — optional mode that presses Enter after pasting (great for chat apps)
  • Multi-monitor overlay — recording indicator appears on every connected display
  • Output recording — capture and transcribe system audio (e.g. meetings)
  • STT providers — OpenAI (gpt-4o-transcribe) or ElevenLabs (Scribe v2), with automatic fallback
  • Self-hosted models — point to any OpenAI-compatible server (whisper.cpp, faster-whisper, LocalAI, Ollama) for fully offline, private transcription
  • AI post-processing — optional LLM pass to fix technical terms using a custom vocabulary
  • Recording history — save, browse, replay, and re-transcribe past recordings
  • Customizable hotkeys — game-style key recorder, set any combination you want
  • System tray — runs quietly in the background, launch at startup
  • Dark & light themes

Requirements

  • Windows 10/11, macOS 12+, or Linux (X11 — requires xdotool)
  • An API key from OpenAI and/or ElevenLabs, or a self-hosted OpenAI-compatible STT server

Installation

Download the latest build for your platform from Releases:

Platform Format
Windows .exe installer (NSIS)
macOS .dmg (x64 & Apple Silicon)
Linux .AppImage or .deb

On first launch, open Settings from the system tray and enter your API key.

Linux note: Auto-paste requires xdotool. Install it with sudo apt install xdotool (Debian/Ubuntu) or your distro's package manager.

Development

# Install dependencies
npm install

# Run in development mode
npm run dev

# Run tests
npm test

# Type-check
npm run typecheck

# Build installer (pick your platform)
npm run build:win
npm run build:mac
npm run build:linux

Project structure

src/
  main/           Electron main process
    dictation/      Hotkey state machine, overlay windows, auto-paste
    transcribe.ts   OpenAI & ElevenLabs STT with fallback
    settingsManager.ts   Settings persistence
    recordingStore.ts    Audio file storage
    errorHandler.ts      Error categorization & notifications
  renderer/       React UI
    components/     Settings form, overlay, recordings panel
  preload/        IPC bridge between main & renderer
tests/            Vitest unit tests

Tech stack

Self-hosted / offline mode

Whisperio can work fully offline with any OpenAI-compatible STT server. In Settings > Providers, set:

  • API Base URL — your server (e.g. http://localhost:8080/v1)
  • Model — model name your server expects (e.g. whisper-large-v3)
  • API Key — leave empty if your server doesn't require one

Compatible servers:

Server Install
whisper.cpp ./server -m ggml-large-v3.bin
faster-whisper-server pip install faster-whisper-server && faster-whisper-server
LocalAI docker run -p 8080:8080 localai/localai
Ollama ollama serve (with whisper model)

Settings

Setting Description
STT Provider OpenAI or ElevenLabs
API Base URL Custom endpoint for self-hosted models (leave empty for official OpenAI)
Model Whisper model name (default: gpt-4o-transcribe)
Fallback Auto-switch provider on failure
AI Post-Processing LLM corrects technical terms after transcription
Custom Vocabulary Comma-separated terms for better recognition
Transcription Prompt Guide the STT model's output style
Dictation Hotkey Global shortcut to start/stop recording
Dictate & Send Hotkey Same as above, but presses Enter after paste
Output Recording Hotkey Record system audio instead of microphone
Save Recordings Keep audio files for later review
Launch at Startup Start Whisperio when Windows boots
Theme Dark or light

License

MIT

About

Global dictation app for Windows, macOS, and Linux. Press a hotkey, speak, auto-paste. Cloud or fully offline with local Whisper models.

Topics

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Sponsor this project

 

Packages

 
 
 

Contributors