Settings

Accessing Settings

Click the Parrot Scribe menu bar icon and select Settings....

General

Language

• Interface Language: Use the macOS system language or choose a supported language.
• Speech Language: Choose Auto (Multilingual) or English Only. English Only can use specialized English transcription models on lower-powered systems.
• Expected Languages: When Speech Language is Auto, select the languages you expect to improve multilingual accuracy.

Recording and startup

• Show notifications: Show a notification when recording starts and stops. Use the adjacent button to open macOS notification settings.
• Launch at Login: Start Parrot Scribe when you log in.
• Auto-capture on Launch: Start recording and transcribing as soon as Parrot Scribe opens.

Speakers

• Auto-show speaker identification: Open Identify Speakers when a recording with unknown speakers ends. When disabled, the session remains available in Library.
• Prompt for speaker consent: Ask before storing a new speaker's voice profile.

Sources

Choose which audio sources Parrot Scribe captures. You can enable either source or both.

• Transcribe Microphone: Capture your voice and room audio from the macOS default input device.
• Transcribe System Audio: Capture audio playing through your Mac.

Transcription

Choose Simple for named presets or Advanced for direct tuning.

Simple mode provides:

• Speech Detection: How readily Parrot Scribe treats incoming audio as speech.
• Responsiveness: The balance between update speed and transcription stability.

Advanced mode exposes the speech detection threshold, transcription interval, audio overlap, and related timing controls. Use it only when the Simple presets do not fit your environment.

Output

• Output Format: Choose the transcript file format.
• Transcript Retention: Set how long complete sessions remain in Library. Expiring a session also removes its retained audio. Tightening retention shows the affected files before you apply it.
• Audio Retention: Retain encrypted audio for playback and post-session speaker identification. Limit retained audio by time or storage size.

See Audio Retention for encryption, deletion, and key recovery details.

Models

• Transcription Model: Download, select, or remove Whisper models. Larger models generally improve accuracy but need more storage and processing power.
• Speaker Identification Model: Download or remove the model used to identify speakers when audio retention is enabled.

Shortcuts

Assign global shortcuts for Start Listening, Stop Listening, New Session, and Open Library. Click a shortcut field, press the new key combination, then click outside the field.

MCP

The MCP server lets compatible local AI clients access Parrot Scribe through explicit per-client permissions. It is disabled by default and requires Pro.

Enable the server only when needed. Create one token per client, copy it when shown, and set that client's access level to Disabled, Enabled, or Trusted. live_transcript is a best-effort in-memory ring that is lost on restart; its sinceSequence value is not a crash-safe cursor. See MCP Server for setup and authorization behavior.

Updates

• Update Channel: Choose Stable for production releases or Latest to include prereleases.
• Automatic Updates: Check manually, notify when an update is available, or download and install automatically.

License

Personal use does not require a license. The License pane shows an active Pro, organization, or legacy license and provides device controls when the license supports them.

See Activation and Deactivation for activation, device management, and transfers.