The theme of this release: transcripts you can trust and correct.
New
- Edit transcripts. Rename any speaker across the whole transcript, and fix any misheard line inline (hover a transcript turn and click the pencil). Corrections flow through to exports and to the summary the next time you regenerate it.
- Audio playback with click-to-seek. Play the recording from the Summary page and click any transcript timestamp to jump straight to that moment.
- Consent, visible after the call. Every recording now shows its consent status in the Library, and recordings without a confirmed consent record (including uploads) can be attested after the fact. Consent disclosure scripts are editable in Settings.
- A sample interview on first launch. New installs open to a completed sample recording, so you can see the full transcript and summary before your first real call.
- Update checks are now optional. A new toggle in Settings > Privacy & Reporting turns off automatic update checks entirely; the manual check button still works.
Improved
- Summaries render as clean formatted text instead of raw markdown.
- Transcription language is auto-detected by default, with a language picker in Settings.
- Summary safety enforcement: a generated summary that adds judgments not grounded in the transcript is automatically regenerated or replaced with a factual fallback. Your summaries only say what the conversation supports.
- Clearer error messages throughout, with no developer jargon.
Fixed
- Quitting mid-transcription no longer strands the recording. Work resumes automatically on next launch.
- Stuck jobs are detected within minutes and retried or reported honestly, instead of spinning forever.
- Swapping speakers now also corrects the text used for summary regeneration, so summaries always reflect the right person's words.
Install: download the DMG below and drag HireScribe to Applications.
Existing installs on macOS update automatically.
Full changelog: https://github.com/SourceBuddy-ai/HireScribe/blob/main/docs/CHANGELOG.md