Browser extension

Check any page you're building from your browser, and when you sign in, see which of its issues real users hit.

On this page

The Accessible browser extension is coming to the Chrome Web Store. It will be released as open source under the MIT license.

Checks that run in your browser

These need no account. Everything runs on the page in your browser, and nothing is sent to Accessible.

Automated WCAG check. An axe-core check against WCAG 2.2 Level AA, using the same rules Accessible's scans and real user monitoring use.
Screen reader transcript. Each element's role, accessible name and state, in reading order.
Keyboard order and walk. The order Tab visits controls, and a walk you do yourself with real Tab presses.
Form check. Every field with the name a screen reader announces and the barriers on it.
Diff against the last scan. What you fixed and what is new since the previous check of the page.
Clear saved scans. A footer button that deletes the per-page scan history the extension keeps in your browser.
Guide links. A finding with a covered WCAG criterion links Accessible's guide for that criterion (21 criteria). The same link goes into the copied agent prompt.
Free export. Findings as Markdown, CSV, JSON, or a printable report you can save as PDF. Made locally in your browser, with no account.
Page outline. A second view in the Screen reader tab: the headings tree and the landmarks, with flags for skipped heading levels, no h1, more than one h1, no main landmark, and unnamed duplicate landmarks.
Show tab order on page. A checkbox on the Keyboard tab that draws numbered Tab-stop markers on the page. They stay aligned when you resize or scroll.
Copy for an AI agent. The findings, formatted to paste into a coding assistant.

The extension highlights elements on the page and changes nothing else.

Under the results, a link reads Check every page of your site in Accessible when you are signed in and the site is on your account; it opens that site's dashboard page. Otherwise a Start free sign-up link is shown.

AI review (experimental)

AI review is experimental and off by default. Turn it on under Experimental: AI review in the panel; the setting appears only where Chrome reports its built-in AI can run. It uses Chrome's built-in, on-device Gemini Nano model, through the Prompt API, to review the alt text of images and the names of links and buttons on the page (up to 60 items). For each one, the on-device model gets its name, an image's file name, and a short piece of the text around it. Its output is shown as labelled "AI suggestion"s, separate from the axe-core findings.

Requirements. A browser, version and device where Chrome's built-in AI is available. See Chrome's documentation (opens in a new tab) for a general reference.
Availability. The section appears only in browsers that provide Chrome's built-in AI. It is not available in Firefox.
First use. The first run downloads the model once.
Privacy. The review runs on your device. Nothing is sent anywhere.
Reliability. AI results vary from run to run and aren't guaranteed to be accurate. Treat them as suggestions to check, not as findings.

What signing in adds

Signing in with your free Accessible account shows Accessible's open issues for the page and marks the ones this check also found, with the real-user impact each one has measured. If the page's site is not yet monitored, you can turn on monitoring from the panel.

Sign in
Choose Sign in in the panel. Accessible asks you to allow the extension, then returns you to the browser.
Open a page of one of your sites
The panel lists the page's open issues. Sites outside your organization show nothing.
Turn on monitoring
If the site is not monitored, choose Turn on monitoring. The panel shows the snippet to add to your site.

Opening it on a page

Open the panel with the Accessible toolbar button or Alt+Shift+A (change it at chrome://extensions/shortcuts). By default the extension can read only the tab you open it on, so a new tab or another tab you switch to asks you to click the button again. To skip that, choose Allow on all sites in the panel, or Always allow to remember one site. Each tab keeps its results while the panel is open.

Permissions

Permission	Why
`activeTab` and `scripting`	To run the checks on the tab you open the extension on, only when you ask.
`sidePanel`	To show results beside the page (Chrome and Edge).
`storage`	To keep you signed in and remember your last checks, for the diff, in your browser. Delete them with Clear saved scans.
`identity`	To sign you in to Accessible.
`app.accessible.so` (host access)	To read your sites' issues once you sign in.
Optional site access	Never asked for at install. Requested only when you click Allow on all sites, or Always allow (or Allow on this site) for one site, so tabs are ready to check without clicking the toolbar button each time. Checks still run only when you ask.

The extension does not ask to read every site at install. If access to a page is lost, for example after it navigates, the panel asks you to allow it on that site.

What leaves your browser

Signed out, nothing. Signed in, the extension sends Accessible the page's address with your sign-in token, to look up that site's issues. The one change it can make is turning on monitoring for one of your sites, when you click to do so. Nothing else from the page is sent.