Skip to main content
Auto Layouts, in the editor’s Layout panel, analyzes a clip’s transcript and video and adds a sequence of layouts (camera bubbles, screen-only beats, camera cutaways, side-by-side splits, and TV-presenter compositions) in an editing style that fits your content. It works one clip at a time on clips under 20 minutes.

​
How to generate auto layouts

  1. Open a clip in the editor and go to the Layout section. 1 2.png
    On short browser windows, Tella keeps Auto Layouts and Generate visible but hides the optional preview and feedback controls to leave room for layout presets.
  2. Click Generate. Tella picks the editing style based on the content of your clip, so there’s no style to choose first. 3 1.png Layouts start streaming in and are applied to your clip as they generate. 3 2.png
  3. Wait for the generation to complete.
  4. Review the result. Give a thumbs up or thumbs down to rate it. Your feedback helps tune the model. 5 2.png

​
What can I do after generating?

Once auto layouts finishes, you can:
  • Click Regenerate for an alternative take. You can tell Tella what didn’t work well, and it uses your notes in the next generation
  • Delete individual layouts while keeping the rest
  • Adjust the start and end of any layout
  • Undo or redo. You’re never locked in
    • Windows: Ctrl + Z to undo, Ctrl + Y to redo
    • macOS: Cmd + Z to undo, Cmd + Shift + Z or Cmd + Y to redo
  • Rate the result with the thumbs up/down buttons

​
Which editing styles are available?

Tella picks one of these styles for you based on what you say and show. Each style is tuned for a different kind of content and a different canvas, and the styles available depend on your clip’s recording type and canvas size.
  • Demo: high-energy proof + commentary; mixes screen reveals, camera cutaways, and punch-ins for product walkthroughs, page critiques, and feature launches
  • Tutorial: procedural, step-by-step instruction where the viewer needs to follow exact clicks, typing, and configuration
  • Presentation: speaker-led talks, pitches, recaps, and lessons; includes a hosted TV show variant on landscape canvases that frames the speaker like a presenter
  • Intro & Outro: only bookends the clip with a camera-led intro and outro and leaves the middle untouched

​
Image B-roll

Auto layouts doesn’t add AI-generated image B-roll. To cut away to an image or video, add media (B-roll) to any layout after generating.

​
Which recordings and canvas sizes are supported?

Auto layouts works with:
  • Landscape (16:9): two-track (camera + screen), screen-only, and camera-only recordings
  • Portrait (9:16): two-track, screen-only, and camera-only recordings
  • Square (1:1) and round: two-track and camera-only recordings
The available editing styles and compositions adapt to the canvas, for example, portrait clips use a presenter-forward split, and landscape presentations can use the TV-presenter framing.

​
Best practices

  • Start with a clean slate: auto layouts works best on raw, unedited recordings
  • Crop first: if you need to crop your screen, do it before running auto layouts
  • One clip at a time: generate layouts per clip; layouts won’t generate across an entire video in one pass

​
What to expect

  • Variability: each generation can produce slightly different results; regenerate to find the version you like
  • Pacing: overrides are used as exceptions to the base layout, so most of the clip stays on the base composition with timed accents around important moments
  • Processing time: longer clips take longer to generate

​
What are the limits of auto layouts?

  • Maximum clip length: under 20 minutes
  • Each person can start up to 10 auto layout generations per minute in each workspace. If you see Couldn’t load Auto Layouts after repeated attempts, wait a minute and try again
  • Make all crop changes before running auto layouts. Cropping afterwards affects every layout and can produce unexpected results
  • After cropping, you can’t update the generated layouts
Last modified on October 19, 2026