TL;DR
Faster transcription, tighter timestamps, and more reliable exports so you can go from upload to publish-ready clips with less cleanup.
What shipped in this release
- Queue upgrades: Faster processing for mid-length files without sacrificing accuracy.
- Timestamp alignment: Cleaner sentence boundaries for clips and chapters.
- Export consistency: TXT and timestamped formats keep section breaks intact.
Why it matters for creators
When timestamps drift, editors lose time re-syncing captions and scenes. Cleaner alignment means fewer manual fixes and a faster path to shorts, highlights, and chapters.
Before vs after workflow
- Before: Upload, export, fix timestamps, then cut clips.
- After: Upload, export, and cut with minimal cleanup.
How to use the update
- Upload audio in AI transcription.
- Select the primary language or use auto-detect.
- Export TXT or the timestamped format and move directly to editing.
Quick checklist for best results
- Trim long silences before upload.
- Pick the primary language when audio is mixed.
- Use the timestamped export for clips and chapters.
FAQ
Does faster processing reduce accuracy?
No. The speed gains come from queue and pipeline optimizations, not from smaller models.
Do I need to re-upload old files?
No. The improvements apply to new jobs only.
Which export should I use for chapters?
Use the timestamped transcript to keep section boundaries and timecodes intact.
Ready to test the update? Start a job in AI transcription and compare the new timestamps.