What exactly is in the generated prompt?
The recreate brief names the hook archetype, beat order and timing, pacing, tone and visual treatment as directives (e.g. 'open on a bold contrarian claim; 3 rapid proof beats of ~8s; end with a question CTA') — structure only, never the original's wording or topic, so your output is original.
Is it a text-to-video prompt or an analysis?
Both: you get the analysis (scores, hook breakdown, retention risks) and a single ready-to-run generation brief distilled from it. The brief works in ViralMint's generators and in external tools alike.
Does it work on videos without captions?
Yes — the transcript comes from local Whisper speech recognition, not from the platform's captions, so any video with speech works in 100+ languages.
Is it free?
Downloading and transcribing run locally and are free. The AI analysis is a cloud call billed per use from a prepaid balance — no subscription.