Capture and compare

All five testing capabilities below are included on Free in Chrome within 20 completed workflows per month. Firefox 153+ is a new, limited-feature desktop target for capture, A/B and Multi comparison, UI review, and optional axe checks. Responsive device presets and browser locale simulation are disabled in Firefox, including Pro. Existing responsive/localized captures remain reviewable and exportable; no new simulated captures can be made there. See browser scope and store availability.

A completed A/B pair or Multi group counts once. These are visual review and optional page checks, not an unlimited or automated functional-testing service.

Choose what to capture

Open SnapDiff on the source tab and choose Single, A/B, or Multi. Then select the capture area:

Area selection cancels if the page scrolls, resizes, or changes. Full-page capture restores the original scroll position and temporary page styles afterward; independently scrolling secondary panels are not expanded.

Infinite feeds: Full page captures the document or main panel height measured at the start, before scrolling to load images. Newly appended posts below that boundary are excluded. SnapDiff does not keep scrolling to find an end to a never-ending feed. Content that shrinks, shifts, or is replaced inside the capture region can still stop capture rather than save an incorrectly stitched image.

Screenshot-only pages: On the Chrome Web Store and non-HTTP pages such as local files or browser-internal pages, use Viewport, Current tab, Browser default, and turn accessibility checks off. Use screenshot-only switches these options for the current popup without overwriting saved preferences. Capture uses the browser's visible-tab screenshot API without injecting scripts. Browser permissions or policy may still block it. SnapDiff's own editor and extension pages are excluded.

Review UI states

For UI testing, capture the layouts, dialogs, or page states you want to review, then inspect spacing, content, and visual changes. Use Single for one state, A/B for two versions, or Multi for a small set. Add annotations to explain what needs attention.

SnapDiff supports visual UI review; it does not execute automated functional tests or validate behavior on physical devices.

Compare A and B

Visual A/B testing means screenshot comparison. It does not split live traffic, measure conversion uplift, or calculate statistical significance.

1. Choose A/B and capture A - expected.
2. Switch to the actual page, reopen SnapDiff, and capture B.
3. Review both images in A / B view, or choose Diff to highlight pixel differences when available.
4. Annotate either side and export the comparison.

A and B may come from different domains, such as a staging site and a live site. Keep each source tab unchanged until its own capture finishes. Closing the popup does not lose a saved A.

Images use native-pixel, top-left alignment; SnapDiff does not resize originals to make them match. Use the same viewport, zoom, and page state for a meaningful pixel comparison. Different device viewport sizes use side-by-side review without a pixel Diff score.

Collect a Multi group

Choose Multi to collect 2–5 images. Change the source page or capture area between shots; device and language changes are available only in Chromium. Click Finish N captures after 2–4 images; the fifth finishes the group automatically.

Multi is visual review only, without pixel Diff. Images keep their own dimensions, annotations, device and locale labels, and timestamps. Exports combine them into one side-by-side image; shorter images leave blank space below rather than stretching.

Only one A/B or Multi draft can be active at a time. Resume or discard it before starting another group; Single captures remain independent. Unfinished groups can also be opened from Captures. A completed A/B or Multi group counts as one workflow, not one per image.

Try responsive device presets

Chromium only. Disabled in Firefox in both plans. Responsive testing lets you visually review a layout at different emulated viewport sizes. Saved responsive captures remain reviewable and exportable in Firefox.

Open More options → Device, choose a named preset, then use Viewport or Full page. Current tab keeps normal capture behavior and is required for Area capture.

Presets temporarily change the page's responsive CSS viewport and capture at 1x. SnapDiff restores the original viewport, zoom, and scroll position. Close DevTools and keep the source tab visible during capture.

These are Chrome-rendered viewport sizes, not another browser engine, physical device pixel density, touch behavior, user-agent switching, or notch and foldable simulation.

Simulate a browser language

Chromium only. Disabled in Firefox in both plans. Localization testing lets you inspect how a supported site responds to locale and language signals. Saved localized captures remain reviewable and exportable in Firefox.

In More options → Language, choose a locale for each capture. Browser default applies no override and does not reload.

SnapDiff temporarily changes the source tab's language headers and browser locale signals. It does not translate the page or force right-to-left layout. The site must support the language; its URL, cookies, saved settings, or language picker may take precedence.

During its own language reload, SnapDiff follows redirects and additional page loads on the same origin (the same scheme, hostname, and port). It waits for loading and same-origin page/data requests to finish, then checks for a quiet DOM, loaded fonts, and completed visible images. Navigation and content must settle for one second within a 30-second limit. Persistent WebSocket connections are not treated as pending page requests. The capture records the final localized URL. After capture, SnapDiff clears the overrides, returns to the original URL, and restores the scroll position where the refreshed page allows it; the site may apply its own same-origin redirects again.

On live feeds such as LinkedIn, a reload may fetch different posts or reorder content. SnapDiff captures the refreshed feed after it settles; it cannot preserve the exact posts from before the reload. Sites with continuous page updates or data requests may reach the timeout instead of producing a capture.

Switching tabs, leaving the original origin, or navigating after the reload has settled still cancels capture. SnapDiff does not automatically follow a different host, login service, or browser-protected page. If the site changes origins for a language, open that destination first and start a new capture. Restoration resets browser overrides, not cookies or account preferences changed by the website.

Language simulation works independently on A and B and does not by itself disable pixel Diff. Device-size differences still use visual review.

Add an optional accessibility check

Accessibility testing is an optional axe scan to help identify issues, not a comprehensive accessibility audit or certification.

Enable Check accessibility in More options before capturing. SnapDiff runs a locally bundled axe check and saves the report with that image. Open Accessibility in the editor to review findings, affected elements, and items that Need review.

Checks are page-wide even for Area screenshots. They cover the loaded main document and open shadow roots, not embedded frame contents or closed shadow roots. A failed or incomplete scan is not a pass. Older images marked Not checked cannot be audited retroactively.

No findings does not mean the page is fully accessible or WCAG-compliant. Manual keyboard and screen-reader checks are still necessary. Export reports carefully: they can include private URLs and HTML snippets.

Having trouble? See capture and export recovery steps.