Quick start
Install Ditto, pick a model, and you're transcribing in under three minutes.
1. Install
Download Ditto-Setup-1.0.0.exe from the Download page and run it. Ditto installs to your user folder. No admin rights needed.
2. Pick a model
The first time Ditto launches, a Welcome window appears. Pick a Whisper model and Ditto downloads it from HuggingFace to your %APPDATA%\ditto\models\ folder.
| Model | Size | Speed | Use case |
|---|---|---|---|
| Tiny | 75 MB | Fastest | Quick notes, lowest latency |
| Base | 142 MB | Fast | Casual use, balanced |
| Small | 466 MB | Medium | Daily use, better accuracy |
| Medium | 1.5 GB | Slower | High accuracy, more CPU/GPU work |
| Large-v3 | 2.9 GB | Slowest | Maximum quality |
When the download finishes, the Welcome window closes itself and the floating pill appears on screen.
3. Your first transcription
- Click into any text field — a chat, a browser tab, your code editor, anything.
- Press Ctrl+Shift+Space.
- Talk.
- Press Ctrl+Shift+Space again.
A second or two later, your transcription is pasted where your cursor was.