Zellio.io

Screenshot Studio

Frame a screenshot, cover what should not be shared, and export it at the size you need.

Runs locallyWorks offlineShare link carries settings, never your data
Send output toImage RedactorImage CropperImage ResizerBuild a workflow from thisOutput stays on this page until you send it.

Frequently asked questions

Can I give it a URL instead of a screenshot?

No, and not as an oversight. A page in a browser tab cannot read the pixels of a site on another domain — that restriction is what stops any page you visit from photographing your webmail — and services that offer it do the work on their own servers with a browser farm. There is no server here. The practical consequence is a fair trade rather than a loss: you take the capture, so anything you can see works, including the half of the web that sits behind a login.

Is Hide actually safe, or is it just a black box drawn on top?

It is safe in the file. The region is painted over before the PNG is encoded, so the pixels underneath are not in the download — there is no layer to peel back and nothing to recover. That is different from covering something in a document or a design tool, where the original is still in the file under the shape. Blur is the one to be careful with: it transforms the pixels rather than replacing them, and published techniques read text back out of a blur, which is why it is not the default.

Why is my export smaller than the 3x I asked for?

Because the canvas would have been bigger than the browser will allocate. A 4K screenshot with padding at 3x is well over eight thousand pixels on its longest edge, which some browsers refuse outright and all of them are slow at. Rather than fail at the moment you press the button, the scale drops to the largest that fits and the tool says what it used. Smaller captures are unaffected.

Does the preview match the file exactly?

Yes. One drawing routine produces both, run at different scales, so the preview is the export at screen size rather than an approximation of it. The one deliberate difference is the checkerboard behind a transparent background: it is there to show you that the area is transparent rather than white, and it is never drawn into the file.

Can I use my own background image or a logo?

Not yet. Gradients, a flat colour and transparent are what there is. Transparent is the useful one if you have a background already — export with no background and place the framed window on your own artwork in whatever you are building the slide in.

Will it open a screenshot from my phone?

PNG, JPEG, WebP and GIF all work, and so does HEIC — what an iPhone saves by default. That one needs explaining: Safari decodes HEIC through the operating system, while Chrome and Firefox cannot and no setting changes it, so this tool carries libheif compiled to WebAssembly and decodes the picture itself. It is fetched only when you actually hand over a HEIC, costs about 350 KB once, and runs in the tab like everything else — the photograph is not uploaded to be converted. On Safari none of that is downloaded, because the browser has already done the job.

Where do the marks go when I change the frame?

They stay where they were on the picture. Marks are stored as fractions of the screenshot rather than as pixel positions, so changing the frame, the padding, the shape or the export scale moves them along with the thing they are covering. A box over a password that stayed upright while the window tilted away from it would stop covering the password, which is why the marks rotate too.

Pro tips

  • Cover things before you fiddle with the frame. The marks move with the picture, so nothing is lost either way — but the one failure that matters is forgetting, and it is hardest to forget while the private row is still the thing you are looking at.
  • Use 1.91:1 for anything that will be a link preview. It is the Open Graph shape, so the card is shown whole instead of being centre-cropped by whatever reads it.
  • A small tilt reads as depth; a large one reads as a mistake. Three or four degrees is usually enough, and it works best with a shadow rather than instead of one.
  • Export at 2x by default. A 1x image on a retina display looks soft in a way most people notice without being able to say why.
  • Keep the frame off for a picture that is already a window. A browser frame drawn around a screenshot of a browser gives you two address bars and looks like an error.

About Screenshot Studio

The services that do this charge per screenshot because they run a browser farm to visit the URL for you — which also means they cannot reach anything behind your login, and that the page lands on their servers before it lands anywhere else. Starting from a screenshot you already took skips both problems: your dashboards and your inbox work like any other picture, and the file is drawn in this tab.

A screenshot straight off the key is rarely the thing you want to post. It is the wrong shape for the place it is going, it has a hard edge where the window was cut, and it usually contains one row of something — an email address, a customer name, a balance — that was fine on your monitor and is not fine in a public thread. This handles all three in one pass, and the order it works in is deliberate: frame it, shape it, then cover what should not travel, because the covering is the part you must not forget and it belongs last where you can see it against the finished picture.

The frame is chrome drawn around the picture, never a slice taken out of it. That sounds obvious and is the single most common fault in tools like this: compute the padding first, fit a title bar inside the result, and the top of someone's screenshot quietly disappears under the fake address bar. Here the bar's height is added to the picture before any spacing is worked out, so every pixel you captured is still in the file. The bar is also capped, because a title bar scaled as a fraction of a very wide screenshot ends up comically tall.

Forcing a shape only ever adds space. Ask for a square, a 16:9 slide or a 4:5 portrait and the canvas grows on whichever axis is short — never crops to reach the ratio, because throwing away part of a screenshot is the one thing a screenshot tool must not do. A wide shot asked for portrait grows downwards; a tall one asked for landscape grows sideways. The background fills whatever that leaves.

Covering something is where this tool takes a position rather than offering a slider. Hide paints over the region, so the detail is absent from the file you download — there is nothing underneath to recover. Blur is offered because people ask for it and it suits tidying a busy shot, but published methods reconstruct text from a blur, and reaching for blur is usually a sign that someone believes they are hiding something. So the page says which is which at the moment you pick one, and Hide is the default. This is the same position the Image Redactor takes, on purpose: two tools on one site must not make different promises about the same risk.

Numbered steps renumber themselves. Place a fourth circle between the first and the second and it becomes the new second, because that is what you meant — the numbers run down the page and left to right across a line, not in the order you happened to click. Keeping insertion order instead would number a how-to 1, 3, 4, 2 and ruin the instructions the picture was made to give. Every mark is stored as a fraction of the picture rather than in pixels, which is why they stay correct when you change the frame underneath them and why they land in the right place at three times the size.

Export is a PNG at one, two or three times. Two is the useful default for anything that will be viewed on a laptop with a retina display, where a 1x image looks soft. Very large screenshots are the exception worth knowing about: a 4K capture at 3x with padding is past the canvas size some browsers will allocate, so rather than failing at the moment you press the button, the scale steps down to whatever fits and the tool tells you what it settled on. There is also a copy button, which puts the finished picture on the clipboard so it can go straight into the post.

Getting a screenshot in is a paste. Press the key, press Ctrl+V on this page, and it is there — saving it to disk first is a step the tool can simply skip, and it is how a screenshot actually arrives. Dropping a file and choosing one both work too. Everything after that is a canvas in your own tab: no upload, no account, no queue, and no limit on how many you make, because there is no server whose time has to be rationed.

Common use cases

  • Posting a product screenshot to a thread or a changelog with a background and a window frame, instead of a bare rectangle with a hard edge.
  • Sharing something from a dashboard, an admin panel or an inbox — pages a screenshot service cannot reach, because they are behind your login and not its.
  • Writing a how-to where the steps need numbering on the picture, and inserting a step later without renumbering the rest by hand.
  • Covering a customer name, an API key or a balance before a screenshot goes into a ticket, a deck or a public issue.
  • Making an Open Graph or social card out of a screenshot at exactly the ratio the platform crops to, rather than letting it choose.
How it comparesScreely, Shots.so, Pika and CleanShot all do the framing well, and the paid ones add cloud history, teams and video. ScreenshotOne and its kind solve a different problem entirely: give them a URL and their servers visit it for you, which is what you want for ten thousand pages from a script and is the one thing this cannot do — a browser cannot read another site's pixels, and there is no render server here to do it instead. The trade runs the other way too. Because the capture is yours, pages behind a login work, bot protection never sees a robot, there is no per-shot cost or queue, and the page you are capturing is never sent anywhere. Automation at scale, theirs; your own screens, privately, this one.