Make an Instagram Story Image

How it works

Your reference image goes to an image-editing model with an instruction describing a vertical Story composition: a strong subject, safe margins for interface overlays, a clean editorial layout, no generated text. A second instruction rules out distorted subjects, unreadable text, watermarks, bad crops, and clutter.

A new picture is generated at exactly 1080 by 1920 pixels. Both numbers are multiples of eight, so nothing needs correcting after the fact. Because the image is composed rather than cropped, a horizontal source becomes a vertical picture instead of a slice of one, and a second run gives a different result.

The app draws on top of your image

A Story fills the screen edge to edge, and the interface is not beside it. It is over it.

Along the top sits your profile picture, your handle, the time posted, and the segmented progress bar showing how many Stories are in the sequence. Along the bottom sits the reply field, the share and like controls, and on some accounts a link or a swipe prompt. None of that is optional and none of it moves.

Between them is a band roughly three quarters of the height, and that band is the only part of your image you actually control. A subject placed with generous headroom in a normal photograph ends up with its head behind a progress bar.

It gets worse with stickers. A poll, a question box, a music tag, or a location pin each occupy a large opaque region wherever you drop it, and those are the things Stories are usually posted for. If you plan to add one, plan its space before you generate rather than after.

One known limitation of the generation itself: the model does not always fill the tall frame. A review on 13 August 2026 returned the subject composed in a square in the middle with solid black bands above and below — a letterbox rather than a full-bleed Story. It is not a failure you can miss, and the fix is to regenerate; a 9:16 frame is a shape the model composes into less reliably than a square. Check the top and bottom of the frame before posting, because those bands sit exactly where the progress bar and reply field go.

Nothing in the picture is tappable

This is the misunderstanding that produces wasted Stories.

The screen is divided into invisible tap zones. A tap on the left goes back, a tap anywhere else goes forward, and a long press pauses. Those zones cover the entire image. Whatever you draw on it — an arrow, a button, the words tap here — is decoration painted underneath a control surface that ignores it.

The only interactive elements are the ones Instagram itself provides, added after upload: link stickers, poll buttons, product tags, the reply field. A picture of a button does nothing, and pointing at one trains people to tap, which advances past you.

So an image that depends on being clicked is an image that fails. Design it to work as a thing seen, and attach the interaction with a real sticker.

Designing for five seconds

The other constraint is time, and it is unusually strict.

A Story advances automatically. Unless someone actively holds the screen, your image gets a few seconds and then it is gone, viewed by a person moving through a queue of them rather than examining any one. There is no scrolling back to reread, no lingering, no second pass.

That rules out a category of composition entirely. Two competing subjects, a detailed scene the eye has to search, small elements that reward inspection, anything requiring the viewer to work out what they are looking at — all of it needs time that the format does not give.

What survives is one subject, large, with a clear separation from its background, resolving in a single glance. The instruction asks for a strong subject and a clean editorial layout for precisely this reason, and it is the part worth judging yourself: look at your generated image for two seconds and then look away. Whatever you can still describe is what the Story communicates.

Publication gate

This page ships once the workflow has been run against a vertical product shot, a horizontal interior, and a portrait, with each output measured at 1080 by 1920 and viewed as a live Story on a phone to confirm what the overlays cover.