Support

The user guide covers day-to-day use. OneSentence is a personal project; there is no support queue or SLA.

Getting started

1. Install a build from GitHub Releases or build from source.
2. On first launch, download AI models (about 1 GB for Parakeet and speech). Whisper is optional; download a model from Settings when you want it. After that, dictation and playback work offline.
3. Look for the waveform icon in the menu bar.
4. Grant Microphone access for Listen and Accessibility access for Read Selection, Snippets, and inserting transcripts. Default recording limit is 300 seconds.

Keyboard shortcuts

• ⌘5 — start or stop Listen (transcript goes to the clipboard and, by default, into the active field).
• ⌘6 — read the current English text selection.
• ⌘⇧Space — open the snippet palette.

Shortcuts can be changed in Settings.

System requirements

• macOS 13.3 or later
• Apple Silicon (M1 and later)
• About 1 GB free disk space for the default models (Whisper is optional)
• Network only for first-run model download and Sparkle updates

Common issues

App not recording

Enable Microphone permission in System Settings → Privacy & Security → Microphone.

Accessibility permission

Read Selection uses Accessibility APIs to capture highlighted text. Enable OneSentence in System Settings → Privacy & Security → Accessibility.

Model download stuck

Check the network connection and free disk space. Retry from the setup screen or Settings. Models come from snowy-0wl/onesentence-models.

Contact

Please do not use GitHub for feature requests or complaints. If something is genuinely broken, you can still open a GitHub Issue or email support@onesentence.app — I may not reply.