Available Now

Tacet

Voice to text that just works. Locally, privately, your way.

GitHub
Tacet - Voice to text that just works. Locally, privately, your way.

Key Features

check_circle

Multi-Engine Transcription

Choose between fully offline local AI, OpenAI API, or Deepgram API. Switch engines anytime based on your needs for privacy, speed, or accuracy.

check_circle

Voice-Powered Productivity

Go beyond basic dictation with voice commands, word replacements, templates, timestamps, and auto-punctuation. All configurable to your workflow.

check_circle

Works Everywhere You Type

Tacet types directly into any active window. Your editor, browser, email, or chat app. No clipboard tricks, no copy-paste, and no app-specific plugins needed.

contact_support

Need help with Tacet?

Contact our dedicated support team for assistance.

Contact Support →

What is Tacet?

Tacet is a speech-to-text app that listens to your microphone and types what you say directly into whatever window you’re working in. Your text editor, browser, email, Slack, whatever. No clipboard, no copy-paste, no app-specific plugins.

It runs fully offline using local AI models, or connects to OpenAI and Deepgram if you prefer cloud. You pick.

The name comes from the musical term tacet, which means “be silent.” It’s the direction on a score that tells a musician to wait, then come in.

What you get

  • Real-time transcription with live preview
  • Multiple transcription engines: local (faster-whisper), OpenAI API, Deepgram API
  • Voice commands you can customize
  • Word replacements for abbreviation expansion
  • Templates for reusable text blocks
  • Timestamp insertion
  • Session history and usage statistics
  • Multi-language UI (English, Spanish, French, German)
  • System tray integration
  • Cross-platform: Windows, macOS, Linux
  • Fully offline mode with no data leaving your machine

Transcription engines

EngineTypeProsCons
Local (faster-whisper)OfflineFree, private, no API key neededNeeds model download, slower on CPU
OpenAI APICloudHigh accuracy, fastNeeds API key, costs money
Deepgram APICloudFast, real-time optimizedNeeds API key, costs money

Voice commands

Voice commands let you trigger actions by speaking. A few examples:

  • Say “new line” to insert a line break
  • Say “period” or “full stop” to insert punctuation
  • Say “delete that” to remove the last transcribed chunk

You can add, remove, or change these in Settings.

Getting started

git clone https://github.com/revoltbots-com/tacet-voice.git
cd tacet-voice
cp config.example.json config.json
pip install -r requirements.txt
python main.py

Supported languages

The UI is available in English, Spanish, French, and German. The transcription language is configured separately and depends on the engine.

License

MIT License. Copyright 2025-2026 Revolt Bots LLC.