Full-page screenshots that do not break, and how to redact them

Chrome's hidden full-page capture, where it fails (sticky headers, lazy images, inner scrolling), and how to mark up a screenshot and remove private details properly before you share it.

By Swapnil Gupta · · 6 min read

A screenshot of the whole page is the quickest way to show a bug, a design review or a before-and-after. Browsers can take one without an extension, but the option is hidden, it breaks on many modern pages, and the image usually needs marking up, and private details removed, before anyone else should see it.

The built-in way, in Chrome and Firefox#

Chrome and Edge have it in DevTools, not in any menu:

  1. Open DevTools.
  2. Open the Command Menu with Ctrl+Shift+P (⌘⇧P on a Mac).
  3. Type "screenshot" and choose Capture full size screenshot. The PNG goes to your downloads. The same menu has Capture node screenshot, for one element.

Firefox has it on the page: right-click, choose Take Screenshot, then Save full page.

Where full-page capture breaks#

A full-page capture has to pretend the window is as tall as the page. Modern pages often react to that, and the image comes out wrong:

  • Full-height sections stretch. A hero set to 100vh becomes as tall as the whole page, because the "viewport" now is.
  • Lazy images are missing. Images that load as you scroll may still be blank, because nothing scrolled.
  • Inner scrolling is cut off. If the content scrolls inside a panel, not the page, only the visible part of the panel is captured.
  • Sticky and fixed parts land oddly. Headers, chat buttons and cookie banners appear once, wherever they happened to be, or cover content.
  • Animations are mid-way. Anything that fades or counts up is caught between frames.

The fixes are practical: scroll to the bottom and back first so lazy images load, close banners and chat widgets, and when a page fights you, capture the visible screen or one element instead. A crisp capture of the part that matters is more useful than a broken image of everything.

Marking it up for a bug report#

A screenshot without marks makes the reader hunt. Circle the problem, add an arrow and a word or two, and say what you expected. With Nitpique, Capture on the bar takes an element, an area, the visible page or the full page, and setting Then to Annotate opens it straight in the Annotator for arrows, shapes, text and highlights. Turn on Bug report and the capture also copies a Markdown report with its measurements, ready to paste into an issue.

Blur is not redaction#

Before a screenshot leaves your machine, look for emails, names, order numbers, tokens in URLs and anything in other tabs or notifications. Then hide them properly.

The same line captured, blurred and redacted. Blur hides the text from a glance but keeps its shape and length; redaction replaces the pixels.

A light blur or a pixelated block keeps the outline of the text, and pixelated text has been recovered by tools built for it, such as Depix. Use a solid block for anything that matters. In the Annotator, press X and drag: redaction removes what was under it from the saved image, so it cannot be undone by anyone you send it to. Keep B (blur) for things that only need to stop distracting, like a busy background.

Cropping is redaction too. If the bug is in the header, a capture of the header alone cannot leak what is in the footer.

A capture-to-share workflow#

  • Close banners, chat widgets and notifications; scroll once so images load.
  • Capture the smallest thing that shows the problem: an element, an area, then the page.
  • Mark the problem with one arrow or box and a short note.
  • Redact personal and private details with solid blocks, not blur.
  • Check the image once more at full size before you share it.
  • Paste it with the page address, the browser and what you expected to see.

Read next