Aspect ratios and safe zones for TikTok, Reels, Shorts and Meta

The same ad runs in several places, and each place wants a different shape. Get the shape wrong and the platform letterboxes the video, crops it, or shrinks it into a corner of the screen. Get the shape right but ignore the interface, and the call to action ends up underneath a caption, a username and a row of buttons.

This guide covers the shape each common placement wants, the parts of the frame the apps cover, and how to plan one shoot or one generation that cuts cleanly to all of them.

The shapes

Export at least 1080 pixels across. Upscaling a smaller video on the platform's side softens it, and on a phone held at arm's length the softness shows first in text.

Safe zones: where the interface covers your video

A vertical video is never seen alone. The app draws its own interface on top: the account name and caption along the bottom, like and share buttons down the right, a progress bar, and on Stories a header at the top. Anything important placed under them is, for the viewer, not there.

The platforms publish their own safe-zone guides, and they change as the apps change. As a working rule that holds across all of them:

• TikTok, Reels and Shorts: keep text, faces and the product out of the bottom 20% and the right 15% of the frame.
• Stories: keep the top 14% and the bottom 20% clear.
• Feed (4:5 and 1:1): the frame is mostly clear, but the ad's own headline and button sit below it, so do not repeat them inside the video at the very bottom.

These margins are deliberately cautious. Before a large spend, check the current safe-zone overlay each platform publishes for its ad placements; ads sometimes carry more interface than organic posts.

The hardest-working pieces of an ad are exactly the ones most likely to land in a covered area: the captions, the price, and the call to action. Place them first, and fit everything else around them.

Captions are part of the frame

Feed video often starts without sound, and plenty of people keep it that way. So the words need to be on screen, not only in the voiceover:

• Burn the captions into the video rather than relying on the platform's automatic ones, so you control the type, the size and the position.
• Keep them above the bottom safe margin, which usually means around the lower third of the clear area rather than the bottom edge.
• Keep each card short: seven words at most, a new card every two or three seconds. Our guide to the anatomy of a 30-second UGC ad covers the word budget in full.

One shoot, every shape

Cropping is geometry, and geometry is unforgiving. A 9:16 frame cut from a 16:9 one keeps only about a third of the original width. A 4:5 frame cut from a 9:16 one keeps about 70% of the original height. Plan for the crops before anything is filmed or generated:

• Frame for vertical first. 9:16 is the shape of TikTok, Reels, Shorts and Stories, and it is the hardest shape to recover from a landscape original.
• Keep the action in the middle. Put the subject and the product in the centre of the vertical frame, so a crop to 4:5 or 1:1 loses only the top and bottom, which are the parts the interface covers anyway.
• Leave headroom for text. A clean area above the subject gives captions somewhere to sit that survives every crop.
• Make landscape separately when it matters. A 16:9 ad for YouTube in-stream is better filmed or generated as 16:9 than cropped out of a vertical original, which would leave it with too little height to hold a face and a product together.

With AI video

The same rules apply to generated video, with one difference: you choose the shape before the shot exists rather than after.

Not every video model renders every shape. Some offer 16:9 and 9:16 only; others offer more. On a Vessero board, each video card offers the shapes the model you pick supports, and the price on its Run button follows the model, length and resolution you choose. When a placement needs 4:5, generate 9:16 and crop, keeping the subject centred, rather than generating 16:9 and losing most of the width.

Keep the product consistent across shapes by generating a key frame first and wiring it into each clip as its first frame, so every version opens on the same product, in the same light.

A checklist before export

1. Each placement has its own export in its own shape; nothing is letterboxed.
2. At least 1080 pixels across.
3. Captions burned in, inside the safe area, seven words a card at most.
4. The price, the offer and the call to action clear of the bottom 20% on vertical placements.
5. The first three seconds work without sound.
6. The last frame is the product, not the presenter.
7. Any AI presenter is labelled wherever the platform expects it — see AI presenters and the FTC rule.