AI delivery workflow · Verismo 1.15
Record → edit the words → captions → ready-to-publish copy.
AI is useful when it shortens the path to a deliverable. Verismo keeps the recording, transcript edit, captions, and Publish Kit on one project timeline instead of scattering them across separate tools.
One delivery pass
The transcript is part of the timeline
Generate it once, then use the same timed words to edit footage, render captions, and ground the copy that accompanies the final video.
- 01
Record the walkthrough
Capture a screen, window, app, or region with narration. The editable project exists on disk before capture begins.
- 02
Select words to cut the footage
Request a Whisper transcript with your own key. Select spoken words and cut or restore their matching video ranges without changing the source recording.
- 03
Turn the transcript into captions
Choose Clean, Boxed, Outline, or Active Word styling. Burn captions into the video or deliver SRT and VTT sidecars.
- 04
Generate the page around the video
Publish Kit drafts a YouTube title, description, and chapters from the transcript and your markers. Chapter timestamps are mapped to the final output after cuts.
A visible privacy boundary
Nothing happens until you ask.
Recording, editing, rendering, and export stay local. AI actions are optional, use your OpenAI API key, and send only the material needed for the action you explicitly invoke.
Read the complete privacy model →The next deliverable
Recording → SOP
The next differentiation is to turn a captured workflow into step-by-step operating instructions anchored to frames and timestamps: a reusable procedure beside the finished walkthrough.
What ships today: video, captions, and publish copy. SOP generation is a product direction, not a current Verismo 1.15 feature.
Questions
What the AI does — and does not do
Does Verismo automatically make editing decisions?
No. You request transcription, select the words to cut or restore, choose whether captions are enabled, and review the generated publish copy. Transcript cuts are non-destructive.
What leaves my Mac when I use AI?
Narration audio is sent directly to OpenAI only when you request transcription. Transcript text is sent only when you explicitly invoke refine, translate, or Publish Kit. Verismo uses your own OpenAI API key.
Do chapter timestamps account for cuts?
Yes. Publish Kit maps transcript segments and timeline markers onto the output timeline, so chapters match the final edited video.
Can Verismo generate an SOP from a recording?
Not in 1.15. Recording-to-SOP generation is a product direction, not a currently shipped feature.