- How does the AI pick the clips?
- The full transcript is analyzed for self-contained moments with a hook — strong claims, stories, emotional peaks, punchlines, surprising facts. Boundaries snap to sentence starts and ends so clips never begin mid-thought.
- What kind of videos work best?
- Talking content: podcasts, interviews, streams, lectures, vlogs. The picker is transcript-driven, so videos without speech aren't suitable — for music or b-roll, use our resize and caption tools directly.
- How is the video cropped to vertical?
- Center-cropped to 9:16 at 1080×1920. Framing that keeps the speaker near the middle of the shot (as most talking-head content does) crops cleanly.
- Can I turn the captions off?
- Yes — toggle them off before processing. On is recommended: most short-form viewing happens muted, and burned captions are the single biggest watch-time lever.
- How long can the input video be?
- Up to your plan's upload limit. Longer videos take proportionally longer to transcribe and render; a 30-minute podcast typically finishes in a few minutes.