July 31, 2026 · 10 min read
App Store screenshot examples: what 10 top listings actually do
I opened ten top-charting App Store listings and recorded every design decision. Nine put the caption on top. Seven give the first screenshot a different job.
Ten top-charting App Store listings, opened and read in one sitting. Every set is portrait. Nine of the ten put the caption above the artwork. Nine keep every caption to seven words or fewer. Seven give the first screenshot a different job from the rest of the set. Six wrap the app in a phone frame of some kind, which means four don't bother.
None of that is a rule Apple enforces. It's where teams with a lot to lose have quietly converged, and the places they disagree tell you which decisions are genuinely yours.
How I looked
On 31 July 2026 I opened ten US App Store listings on the web and recorded what each screenshot set does: caption position and length, whether there's a device frame, what the first image is doing, and what the status bar says. Chart positions come from the CHART field on each listing, read on the same page load.
The sample: ChatGPT (Productivity #1), Spotify (Music #1), Strava (Health & Fitness #1), Duolingo (Education #2), Instagram (Photo & Video #2), Things 3 (Productivity #3), Canva (Photo & Video #4), Notion (Productivity #37), Calm (Health & Fitness #39), and Bear, an Editors' Choice indie app with no current chart placement.
That last one is deliberate. Nine listings from teams with marketing departments tell you what a big budget converges on. One from a small studio tells you how much of it you can do without one.
What Apple actually promises
Worth getting this right, because a lot of advice overstates it. Apple's own product page guidance says you can feature up to 10 screenshots, and that "depending on the orientation of your screenshots, the first one to three images will appear in search results when no app preview is available".
That sentence carries three qualifiers. The count is one to three, not always three. It depends on orientation, so a landscape set gets fewer slots than a portrait one. And it only applies when you have no app preview video, because the video takes the slot when you do.
Apple's guidance for the rest is one line: "Focus each subsequent screenshot on a main benefit or feature so that you fully convey your app's value." There's also a nudge to include at least one Dark Mode screenshot if your app supports it.
So the first one to three images are your search-results card, and everything after that is for people already on your page. Every pattern below is a consequence of that split. Our App Store screenshot sizes reference has the dimensions for each display class.
Pattern 1: the caption goes on top
Nine of ten put text above the artwork. Duolingo is the only set that does the opposite, dropping a solid colour block across the bottom quarter with three or four lowercase words in it.
Top placement makes sense once you picture the search results card. Screenshots there are small and often partly cut off at the bottom of the row. Text at the top is the part guaranteed to survive. Text at the bottom is the part that gets clipped first.
Duolingo gets away with the inversion because its captions are so short and the colour blocks are so loud that they read as bands of colour before they read as words. That's a bet on brand recognition rather than on legibility at thumbnail size.
Pattern 2: the first screenshot does a different job
This was the strongest signal in the sample. Seven of the ten treat image one as something other than "feature number one".
Three different approaches show up. Things 3 and Instagram open with a plain, uncaptioned screenshot: no frame, no background, no headline, just the app. Strava and Notion open with a card carrying no product UI at all, one a solid brand colour with a three-word promise, the other an illustration. Spotify, Canva and Calm keep a normal captioned shot but add branding above it, in Calm's case an Apple App of the Year laurel.
Only ChatGPT, Duolingo and Bear treat image one exactly like the rest.
The common thread across all seven: image one answers "what is this and whose is it", and image two onward answers "what can it do". If your first screenshot is a settings pane with a caption reading "Powerful customisation", you're spending your best slot on your fourth-best idea.
Pattern 3: device frames are a brand decision
Six use a frame, four don't, and the six don't agree on what a frame is.
Things 3, Spotify, Calm and Notion use realistic phone hardware, drawn accurately down to the Dynamic Island or notch. Strava draws a thin orange outline in its brand colour, more diagram than device. Bear uses a simplified bezel with the mascot peeking out from behind it. Duolingo, Canva, Instagram and ChatGPT show the app's own interface with rounded corners and nothing else.
There's a rough split by content. Apps whose screens are mostly text and lists tend to keep the frame, which gives the eye an object to hold. Apps whose screens are already full of images (Instagram, Canva) drop it, because a bezel would only shrink artwork that's doing the selling.
One variation is worth stealing. Calm angles its phones in perspective rather than sitting them flat, which lets two devices overlap without either being hidden. If you want that, our editor exports 3D quarter-turn device views alongside the flat ones.
Every framed set in the sample also runs the device off the bottom edge of the canvas rather than fitting the whole phone inside it. Fitting the whole device in makes it small. Cropping it makes the screen big, and the screen is the product.
Pattern 4: captions are shorter than you think
Nine of the ten keep every caption to seven words or fewer. Most sit between two and five. Notion's run four words each and end with a full stop. Bear's are two to four. Strava opens with three.
Things 3 is the single exception, and it's instructive rather than sloppy: it uses a bold opening phrase followed by a longer plain-weight sentence, so the eye still lands on three or four words and the rest is there if you're interested. Spotify does a version of the same trick in all caps, bolding the two or three words that carry the meaning.
That two-weight caption is the most transferable idea in the whole sample. You get a headline that works at thumbnail size and a sentence that works at full size, in one line of text.
A long caption wastes the slot. At search-card size, a fourteen-word sentence is a grey texture.
Pattern 5: the status bar is a tell
There are four ways to handle the status bar, and my sample used all four.
Things 3 and Spotify show 9:41, Apple's marketing time. ChatGPT holds 11:30 across every shot and Instagram holds 5:26, so both invented a time and kept it consistent. Duolingo crops the status bar out entirely. Notion left real device times in, so its three framed shots show three different clocks at three different battery levels.
Nobody will reject you for this. But an inconsistent status bar across a set is the clearest signal that the images were captured at different moments rather than designed as a group, and it's free to fix. Pick one time, or crop the bar. Our Xcode simulator workflow post covers getting a clean status bar out of the simulator in the first place.
Pattern 6: awards live here, but not on Play
Calm's first screenshot carries an Apple App of the Year laurel. Strava's second is an Apple Watch App of the Year card. Both spend a top-three slot, the most valuable real estate they have, on a third-party endorsement rather than a feature.
Worth knowing that this is store-specific. Google explicitly bans ranking, award and price claims on the Play feature graphic, which we covered in the Google Play feature graphic post. Same company, same app, opposite rule depending on the asset. If you're shipping to both stores, the award badge is an App Store asset that has to come off before the Play upload.
What this sample doesn't prove
Some honest limits, because there's a lot of screenshot advice out there carrying more confidence than evidence.
This is ten listings on one day. It shows what successful apps have converged on. It does not show that these choices caused the success, and I have no conversion data. Big apps also get installs from brand recognition, so they can afford choices that would sink an unknown app, and the reverse is true too.
Category matters as well. Games follow different conventions and aren't represented here at all. Two of the ten are Photo & Video apps, which biases the frame count downward.
Treat the list as defaults you need a reason to break.
Frequently asked questions
How many App Store screenshots appear in search results? One to three, depending on the orientation of your screenshots, and only when you have no app preview video. Apple doesn't publish which orientation gets which count. If you upload a preview, the video takes that space instead.
Should App Store screenshots have text on them? All ten listings I looked at add captions, though Things 3 and Instagram leave their first screenshot bare. Captions aren't required by Apple, but an uncaptioned screenshot has to explain itself using only the app's own interface.
Where should the caption go? Above the artwork, unless you have a specific reason not to. Nine of ten top listings put it there, and the top of the image is the part least likely to be clipped in the search results row.
Do I need a device frame in my App Store screenshots? No. Four of the ten sets use no frame at all. Frames help when your screens are text-heavy and hurt when your screens are already full of imagery. If you do use one, crop it at the canvas edge rather than fitting the whole phone in.
What should my first App Store screenshot be? Something that answers what the app is, not what feature three does. Seven of ten treat the first image differently from the rest, whether that's a plain screenshot, a branded title card, or a normal shot with the wordmark added.
How long should an App Store screenshot caption be? Seven words or fewer covers nine of the ten sets, and two to five words is the norm. If you need a longer sentence, bold the two or three words that carry the meaning so the caption still works at thumbnail size.
Portrait or landscape? Portrait. All ten sets are portrait, and Apple's own wording ties the number of search-result slots to orientation.
Ready to build your set? Open the editor, pick a caption position, and export every App Store and Play Store size from one project. One-time $25, no subscription.