Quick answer
Neither store sets a minimum caption size. Use the thumbnail test: shrink your screenshot to the width it appears in search results, and if you can't read the caption at a glance, it's too small.
The short answer
> Quick answer: Bigger than you think. Neither store publishes a minimum, so the only test that matters is whether the caption reads at thumbnail size.
Shrink your screenshot until it's the width it appears at in search results. If you have to lean in, it's too small.
That one test replaces every rule of thumb you'll read about this.
Shrink your screenshot until it's the width it appears at in search results. If you have to lean in, it's too small.
That one test replaces every rule of thumb you'll read about this.
Why thumbnail size is the real size
Here's the thing most people design past.
You make the screenshot at full resolution on a big monitor, and it looks great, because of course it does.
But the first time anybody sees it, it's a thumbnail in a search result on a phone, competing with four other apps.
That's where the decision gets made. The full size view happens after somebody's already interested.
So the caption isn't there to be read on your monitor. It's there to survive being shrunk.
You make the screenshot at full resolution on a big monitor, and it looks great, because of course it does.
But the first time anybody sees it, it's a thumbnail in a search result on a phone, competing with four other apps.
That's where the decision gets made. The full size view happens after somebody's already interested.
So the caption isn't there to be read on your monitor. It's there to survive being shrunk.
The test, step by step
Two minutes, no tools beyond what you already have.
Export your screenshot. Then scale it down so it's about the width of your thumb.
Look at it for one second and look away. That's roughly the attention a real search result gets.
If you can say what the caption told you, it works. If you can only tell that there is text there, it doesn't.
I ran this on a batch of my own screenshots and half of them failed. Every one had looked fine full size.
Export your screenshot. Then scale it down so it's about the width of your thumb.
Look at it for one second and look away. That's roughly the attention a real search result gets.
If you can say what the caption told you, it works. If you can only tell that there is text there, it doesn't.
I ran this on a batch of my own screenshots and half of them failed. Every one had looked fine full size.
Fewer words is the same as bigger type
This is the trade nobody makes explicitly, and it's the whole game.
You have a fixed area. Every extra word makes every word smaller.
So a three word caption at large type beats an eight word caption that's technically legible.
Cut the sentence, don't shrink the font. That's the rule.
Our caption copy post covers what to cut and what to keep.
You have a fixed area. Every extra word makes every word smaller.
So a three word caption at large type beats an eight word caption that's technically legible.
Cut the sentence, don't shrink the font. That's the rule.
Our caption copy post covers what to cut and what to keep.
Contrast matters as much as size
A big caption in pale grey on white is invisible. Size alone won't save it.
The accessibility standard is worth borrowing here even though the stores don't enforce it. WCAG's contrast rule asks for 4.5 to 1 for normal text.
For large text it relaxes to 3 to 1, where large means at least 18 point, or 14 point bold.
Your caption is large text by that definition, or it should be, so 3 to 1 is your floor.
The reason those numbers exist is practical. The higher ratio compensates for contrast sensitivity loss around 20/40 vision, which is a lot of people.
The accessibility standard is worth borrowing here even though the stores don't enforce it. WCAG's contrast rule asks for 4.5 to 1 for normal text.
For large text it relaxes to 3 to 1, where large means at least 18 point, or 14 point bold.
Your caption is large text by that definition, or it should be, so 3 to 1 is your floor.
The reason those numbers exist is practical. The higher ratio compensates for contrast sensitivity loss around 20/40 vision, which is a lot of people.
The screenshot behind your text is the enemy
Most failed captions fail because of what's underneath them.
App interfaces are busy. Cards, lists, colour, images. Put white text over that and it works in some places and disappears in others.
The fix isn't a heavier font. It's a solid band behind the caption, or a plain area of the canvas the screenshot doesn't reach into.
That's why so many good store listings put the text on a coloured background above the device, not on the device.
It looks less clever and it reads every time.
App interfaces are busy. Cards, lists, colour, images. Put white text over that and it works in some places and disappears in others.
The fix isn't a heavier font. It's a solid band behind the caption, or a plain area of the canvas the screenshot doesn't reach into.
That's why so many good store listings put the text on a coloured background above the device, not on the device.
It looks less clever and it reads every time.
Where the caption sits changes how big it needs to be
Top of the frame, and it's the first thing seen, so it can be smaller and still work.
Bottom of the frame, and it competes with the store's own interface elements in some placements.
Over the device, and it has to beat whatever's on the screen behind it.
So position and size aren't separate decisions. A caption you moved is a caption you need to retest.
Our design mistakes post covers the placements that keep going wrong.
Bottom of the frame, and it competes with the store's own interface elements in some placements.
Over the device, and it has to beat whatever's on the screen behind it.
So position and size aren't separate decisions. A caption you moved is a caption you need to retest.
Our design mistakes post covers the placements that keep going wrong.
What the stores actually constrain
Worth being precise, because the constraint isn't on your type.
Google's asset requirements set a minimum dimension of 320 pixels and a maximum of 3840 for screenshots, with at least two required and up to eight per device type.
That's it. Nothing about text at all.
So the readable caption is entirely your problem, and there's no rejection to warn you when you've got it wrong.
You just quietly convert worse than the app next to you.
Google's asset requirements set a minimum dimension of 320 pixels and a maximum of 3840 for screenshots, with at least two required and up to eight per device type.
That's it. Nothing about text at all.
So the readable caption is entirely your problem, and there's no rejection to warn you when you've got it wrong.
You just quietly convert worse than the app next to you.
Export at the top of the range, not the bottom
A cheap win people miss.
If the maximum dimension is 3840, there's no reason to upload something at 1080 because your design file happened to be that size.
The store scales down for the thumbnail either way, and downscaling from a larger source gives cleaner type edges.
Thin fonts suffer worst from this. A hairline weight at low resolution turns to mush at thumbnail size.
So export large, and pick a weight that survives the shrink.
If the maximum dimension is 3840, there's no reason to upload something at 1080 because your design file happened to be that size.
The store scales down for the thumbnail either way, and downscaling from a larger source gives cleaner type edges.
Thin fonts suffer worst from this. A hairline weight at low resolution turns to mush at thumbnail size.
So export large, and pick a weight that survives the shrink.
Font weight beats font size at small sizes
If you can only change one thing, change the weight.
A medium weight at a given size reads far better shrunk than a light weight two steps larger.
Light and thin weights are designed for large display use. At thumbnail size their strokes vanish.
That's the same reasoning behind the accessibility standard treating 14 point bold as equivalent to 18 point regular.
Weight is doing real work there, and it's free.
A medium weight at a given size reads far better shrunk than a light weight two steps larger.
Light and thin weights are designed for large display use. At thumbnail size their strokes vanish.
That's the same reasoning behind the accessibility standard treating 14 point bold as equivalent to 18 point regular.
Weight is doing real work there, and it's free.
Test it in dark mode too
One test, easy to forget, and it catches a specific failure.
A caption sitting on a light screenshot works. Swap the screenshot to a dark mode capture and the same white caption disappears into it.
If you ship both light and dark screenshots, the caption treatment has to work on both, or you need two treatments.
Our dark mode screenshots post covers whether shipping both is worth doing at all.
If you only ship one, this problem disappears, which is an argument for only shipping one.
A caption sitting on a light screenshot works. Swap the screenshot to a dark mode capture and the same white caption disappears into it.
If you ship both light and dark screenshots, the caption treatment has to work on both, or you need two treatments.
Our dark mode screenshots post covers whether shipping both is worth doing at all.
If you only ship one, this problem disappears, which is an argument for only shipping one.
The first screenshot needs the biggest caption
They're not all equal and they shouldn't be typeset the same.
The first one appears in search results. It gets a fraction of a second and it's fighting for a tap.
Screenshots three through eight are seen by people who already scrolled, on a bigger view, with more patience.
So put your largest, shortest caption first, and let the later ones carry more detail at a smaller size.
Our first screenshot post covers what that slot has to achieve.
The first one appears in search results. It gets a fraction of a second and it's fighting for a tap.
Screenshots three through eight are seen by people who already scrolled, on a bigger view, with more patience.
So put your largest, shortest caption first, and let the later ones carry more detail at a smaller size.
Our first screenshot post covers what that slot has to achieve.
Translated captions get longer
This breaks more listings than any font choice.
German and Finnish run substantially longer than English for the same idea. If your caption exactly fills the width, the translation won't fit.
So design with slack. If your English caption uses the full width, it's already too long.
And never solve an overflow by shrinking the type in one language only. You'll ship one market with unreadable screenshots and never notice.
Shorten the source sentence instead. It helps every language at once.
German and Finnish run substantially longer than English for the same idea. If your caption exactly fills the width, the translation won't fit.
So design with slack. If your English caption uses the full width, it's already too long.
And never solve an overflow by shrinking the type in one language only. You'll ship one market with unreadable screenshots and never notice.
Shorten the source sentence instead. It helps every language at once.
One caption per screenshot
A rule that sounds obvious and gets broken constantly.
Two lines of different sizes on one screenshot means the reader has to decide which one matters, in a fraction of a second, and they won't.
A headline and a subhead works on a web page where the reader has committed. On a thumbnail it's noise.
One idea, one caption, one size.
If you have two things to say, that's two screenshots, and you're allowed up to eight.
Two lines of different sizes on one screenshot means the reader has to decide which one matters, in a fraction of a second, and they won't.
A headline and a subhead works on a web page where the reader has committed. On a thumbnail it's noise.
One idea, one caption, one size.
If you have two things to say, that's two screenshots, and you're allowed up to eight.
How to check it before you ship
A short routine that takes five minutes for a whole set.
Put all your screenshots side by side at thumbnail size on one canvas.
Step back from the monitor until you can't read them, then walk forward until you can. The one that resolves last is your weakest.
Fix that one and repeat.
I counted three rounds to make a set consistent the first time I did this, and one round every time since. You can lay them out on the create page rather than building a canvas by hand.
Put all your screenshots side by side at thumbnail size on one canvas.
Step back from the monitor until you can't read them, then walk forward until you can. The one that resolves last is your weakest.
Fix that one and repeat.
I counted three rounds to make a set consistent the first time I did this, and one round every time since. You can lay them out on the create page rather than building a canvas by hand.
The short version
There's no published minimum, so stop looking for one.
Shrink it to thumbnail size and see whether it reads in a second.
Hold at least a 3 to 1 contrast ratio against whatever is behind it, and put it on a plain area rather than over a busy interface.
Use a medium weight, not a light one, and export at the top of the size range.
Then cut a word. There's always one to cut, and it makes everything else bigger.
Shrink it to thumbnail size and see whether it reads in a second.
Hold at least a 3 to 1 contrast ratio against whatever is behind it, and put it on a plain area rather than over a busy interface.
Use a medium weight, not a light one, and export at the top of the size range.
Then cut a word. There's always one to cut, and it makes everything else bigger.
Frequently asked questions
What size should caption text be on app store screenshots?
Neither store publishes a minimum, so there's no number to comply with. Use the thumbnail test instead. Shrink the screenshot to the width it appears at in search results, glance at it for one second, and if you can't say what it told you, it's too small.
What contrast ratio should screenshot captions have?
Borrow the accessibility standard. WCAG asks for 4.5 to 1 for normal text and relaxes to 3 to 1 for large text, where large means at least 18 point or 14 point bold. Your caption should qualify as large text, so 3 to 1 is the floor to hold.
Should caption text go over the phone screen or above it?
Above it, on a plain area, almost always. App interfaces are busy, so white text over a screenshot works in some places and vanishes in others. A solid band or a plain canvas area reads every time, and it looks less clever for a reason.
Does font weight matter more than font size?
At thumbnail size, often yes. A medium weight reads better shrunk than a light weight two steps larger, because thin strokes disappear first. That's the same logic behind the accessibility standard treating 14 point bold as equivalent to 18 point regular.
How do I stop translated captions from breaking my layout?
Design with slack and shorten the English source rather than shrinking type per language. German and Finnish run longer for the same idea, so a caption that fills the full width in English will overflow. Our [Play screenshot sizes guide](/blog/play-store-screenshot-sizes-2026) covers the canvas you have to work within.
Build the mockup in your browser.
Drop a screenshot into a real device frame and export at the exact store size — free, no signup.