Descript switching to ElevenLabs Scribe V2 transcription is the upgrade that actually changes accuracy on difficult audio
The specific improvement areas are the ones that matter: different accents, audio recorded in less than ideal conditions and proper nouns being correctly identified. All three are where previous Descript transcription produced corrections I had to make manually before the text-based editing workflow became useful.
The 16 transitions with animated previews and bulk application across multiple scenes changes the production workflow for video content. Seeing the transition animation before committing to it eliminates the iterate-and-preview cycle that made transition decisions slow.
The API update enabling triggered web exports and returning URLs for MP4 files is the programmatic output capability that makes Descript usable in automated content pipelines rather than only as a manual editing environment.
For enterprise teams the custom workspaces with granular access permissions are the organisational feature that changes whether Descript is viable for multi-team deployments.
What is your typical audio quality going into Descript? For anyone recording in field conditions or with guests on budget setups, does the Scribe V2 improvement land where you most needed it?