TOOLTIKI Lovable tool, really free

Three shapes cover every feed

Designing separately for every platform is unnecessary. Feed images across all of them come down to three shapes — square, tall portrait, and full-screen portrait — and a fourth, landscape, that survives mainly where video lives.

Design once per shape and the export list gets short. The exact pixel dimensions change from time to time and the shapes have been stable for years.

Why did feeds go vertical?

Because phones are held vertically, and a taller image occupies more of the screen for longer as the feed scrolls past. That is the entire mechanism, and it explains every size change of the last decade.

Feeds settled on a limit somewhere around a 4:5 ratio for in-feed images rather than allowing full-screen height, because a post that fills the whole screen leaves nothing visible above or below it — no author, no caption, no sense that there is more to scroll to.

Full-screen portrait belongs to the formats that are meant to fill the screen: stories, reels and short video. There the whole display is the canvas, and the interface elements sit on top of the image rather than beside it.

Landscape has not disappeared so much as retreated to where it belongs. Video that will be watched on a television or a laptop is still horizontal, and so is anything designed to sit inside a page rather than fill a phone — which is why a thumbnail and a feed image are different jobs even when they show the same thing.

What is a safe area?

The part of the frame nothing will be laid over. A full-screen format has interface on top of it — a profile row at the top, captions and buttons at the bottom — and anything important placed there is covered.

The proportions are roughly consistent across platforms: keep the top and bottom of a vertical frame clear and treat the middle as the usable area. Text placed in that middle band survives every platform without a redesign.

It applies to the sides as well in formats that get cropped rather than letterboxed. A design that runs edge to edge is a design that loses its edges.

How do carousels differ?

Every card has to be the same shape. A carousel picks one ratio from the first image and applies it to all of them, so a set mixing square and portrait gets cropped into agreement — and the cropping is not something you approve.

That makes consistency the first rule and continuity the opportunity. Designing a carousel as one wide image sliced into cards works well, provided the slices are exact and the important content does not sit on a seam.

The first card carries the whole thing. It is the only one guaranteed to be seen, so it has to work alone as well as it works as an opening.

Why does the same image look different on two platforms?

Different compression, different colour handling, and different crops. Each platform re-encodes on upload with its own settings, so the file that reaches a viewer is never the file you sent.

Colour is the subtler difference. An image saved in a wide colour space and displayed by something assuming a standard one comes out flat or oversaturated, which is why exporting in the standard space avoids a class of problem that is hard to diagnose after the fact.

Sharpness suffers most in flat areas and gradients, which is where compression has the least information to work with. A gradient background that looks smooth locally can band visibly after upload.

Text is where the difference shows first. Compression works by discarding detail the eye is least likely to miss, and a hard edge between a letter and its background is exactly the kind of detail it discards — which is why small text on an uploaded image looks fuzzy in a way the same text in a caption never does.

What is worth exporting?

One square, one tall portrait, one full portrait, and a landscape if video is involved. That covers the feeds, and the per-platform differences within each shape are small enough that one export serves several destinations.

Keeping the original at a larger size than any export is the other half. Platforms change their dimensions occasionally, and re-exporting from a master takes a minute where re-shooting does not.

Naming the exports by shape rather than by platform makes them reusable. A file called "square" survives a platform renaming its formats; one named after a product does not.

Questions people ask

Does image size affect reach? Not directly. It affects whether the image is legible on a phone, which affects whether anyone stops.

Should I add a border to fit a square? It works and it costs screen area. A crop that keeps the subject is usually better.

Do stories and reels use the same frame? The same ratio, with different interface over it. Check both safe areas if you post to both.

What about video thumbnails? Landscape for horizontal video, and the thumbnail is a separate design job rather than a frame grab.

Export by shape, not by platform. Social media image sizes lists the current numbers in one place, the social media image resizer and Instagram image resizer do the conversions, and the per-format tools cover Instagram posts, stories, reels and carousels, Facebook posts, LinkedIn, X, TikTok, Pinterest and WhatsApp status.