Cadence
A cross-platform desktop application that provides simple, privacy-focused speech transcription. Press a shortcut, speak, and have your words appear in any text field. This happens entirely on your own computer without sending any information to the cloud.
Windows x64 (Unsigned Beta) · macOS Build Coming Soon
Designed for clarity & control.
A clean, minimal interface that stays out of your way until you need it. Configure models, shortcuts, sound inputs, and paste rules in seconds.
Built to fill the gap for truly open speech-to-text.
Cadence was created to give everyone a lightweight, highly accurate, and extensible voice typing tool without proprietary lock-in.
100% Free & Open
Accessibility tooling belongs in everyone's hands, not behind a paywall. Cadence is free software engineered for daily productivity.
Private by Design
Your voice stays strictly on your computer. Get instant transcriptions without sending audio packets over the network or to cloud servers.
Simple & Focused
One tool, one job. Transcribe what you say and place it into whatever text field has focus — IDEs, web browsers, notes, or email.
How Cadence Works
Seamlessly integrates into your existing workflow across any application.
Press Shortcut
Press a configurable hotkey (e.g. Ctrl + Space) or hold down push-to-talk mode.
Speak Your Mind
Speak naturally while the shortcut is active. Audio recording handles low latency capturing seamlessly.
Release to Process
Release the key. Cadence filters silence with Silero VAD and transcribes with your chosen local AI model.
Direct Text Paste
Transcribed text is pasted immediately into your active window or app with zero manual copying needed.
Powered by state-of-the-art local models.
The entire transcription process happens on your CPU or GPU using industry-standard machine learning primitives.
Voice Activity Detection (VAD)
Silence and background noise are intelligently filtered using Silero VAD before inference begins. This speeds up processing and eliminates unwanted background noise artifacts.
Flexible Model Choices
Choose the model that fits your hardware:
- Whisper Models: Small, Medium, Turbo, Large with GPU acceleration.
- Parakeet V3: CPU-optimized model with exceptional speed and automatic language detection.
Technical Overview
Platform Support
Models & Acceleration
Privacy & Audio
Ready to type at the speed of speech?
Download Cadence for Windows, macOS, or Linux. Free, private, and open source speech-to-text.