Overview
TheTechDad1027 Dictation is a local desktop utility for recording speech, transcribing it locally with whisper.cpp, and placing the finished text on the clipboard for pasting into another application.
Basic Dictation
- Open the Dictate tab.
- Select Start Dictation.
- Speak normally.
- Select Stop Dictation.
- Wait for transcription to complete.
- If Auto-Copy is enabled, right-click and paste the result wherever you want it.
Recent Dictation
The Recent tab stores recent transcriptions locally. Select an entry to load/copy it again. Use the individual × button to delete one item or Clear History to remove the full list.
Settings
- Speech recognition model: Fast (Tiny), Balanced (Base), or Accurate (Small). A model downloads locally the first time it is selected.
- Theme: Light or Dark.
- Microphone: choose the input device used for recording.
- Automatically copy transcription: places completed text on the clipboard.
- Recent dictation entries: controls the maximum retained history.
Information & Logs
The Information window contains About, Help & Information, licensing, and Logs. The Logs tab shows recent server/application activity without requiring you to browse into the Logs/ folder.
Audio Privacy
Dictation is processed locally. Temporary WAV files are used only while whisper.cpp transcribes the recording, then deleted. The application does not intentionally retain or upload those recordings.
Temporary WAV Audio
While processing a recording, the application writes a temporary WAV file under Temp/. whisper.cpp transcribes that file locally. The WAV file is deleted after transcription. Stale dictation temp files from interrupted runs are cleaned up on a later launch.
Local Files
Data/— settings and recent dictation history.Engine/— portable whisper.cpp runtime.Models/— local Whisper model.Logs/— application log.Temp/— temporary transcription working files.Assets/— local branding assets.
Troubleshooting
- If Start Dictation cannot access the microphone, verify microphone permission in Brave and confirm the selected microphone in Settings.
- If transcription fails, open Information → Logs and review the latest entries.
- If the speech engine/model is missing, the application will attempt to download it into its own folder.
- If the clipboard does not update, use Copy Again or verify Auto-Copy in Settings.
- If a recent entry is no longer needed, delete it individually with its × button.
Licensing
The application uses whisper.cpp and OpenAI Whisper model weights under the MIT License. Keep applicable copyright and license notices when redistributing those components.