Skip to content

Signed-in scanning

Account areas, dashboards and checkouts often sit behind a sign-in. To scan them, give the site a session cookie or an authorisation header from a test account. Every scan of the site then loads its pages as that signed-in user.

Signed-in scanning is on every plan.

Credential Use it when What to enter
Header Your site accepts a token on each request, such as Authorization: Bearer …, or sits behind a gateway that reads a header. The header name and its value.
Cookie Your site keeps the session in a cookie after signing in. The cookie’s name, its value, and its domain (for example .example.com). The domain must be your site’s own: its host, a parent domain of it, or a subdomain of it.

You can add up to 10 headers and 10 cookies. A few headers belong to the scanner and can’t be set: Host, User-Agent, Content-Length, Cache-Control, Pragma and Expires.

To find a session cookie, sign in to the site as your test account in a browser, open the developer tools’ storage or application panel, and copy the session cookie’s name, value and domain.

Owners and admins can configure it.

  1. Open the site, then Settings in its sidebar.
  2. In Authenticated scanning, choose Configure (or Replace if credentials are already set).
  3. Add a header (Add header), a cookie (Add cookie), or both, then Save credentials.

The card then lists the names of what is set, never the values. Saving always replaces the whole set; to change one value, enter them all again. Clear removes them, and the site scans signed out.

Then run a scan and open a page that needs the sign-in: its screenshot should show the signed-in page, not your login form.

  • Values are encrypted (AES-GCM) before they are stored. Nobody can read them back, in the dashboard or through the API: the settings show names only.
  • They are decrypted only by the scanner, just before it loads a page, and live in that request alone, never in queues or logs.
  • Each page is loaded in a fresh browser context, so a session never carries over to another scan or another workspace.
  • Headers go only to your site: requests to the scanned page’s own host (with or without www., and from http to https) carry them; requests the page makes to other sites (a CDN, fonts, analytics) do not. Cookies go only to the domain you set, as in any browser.
  • Saving or clearing them is recorded as an audit event with who did it and the header and cookie names, never the values.
  • If credentials can’t be decrypted, the page is scanned signed out rather than not at all; the results show the signed-out page.

Every scan of the site: scheduled and manual full-site scans, single-page scans, deploy webhook scans, Verify now, and the GitHub verification scan after a merge, so a fix is checked in the same state it was found in.

Journeys use them only when the journey’s Use this site’s authenticated-scanning credentials switch is on.

Discovery reads your sitemap without signing in, so pages that only exist behind the sign-in are usually not in it. Add them under Manual URLs (with Discovery mode set to Manual or Hybrid), or turn on Follow links on scanned pages so links found on the signed-in pages are followed.

On public pages the scanner also acts on the page: it hovers over elements, presses single keys, and presses the pointer down without releasing it, to see what reacts. On a signed-in page those actions could do real work in the account (a keyboard shortcut that archives something, say), so a scan that sends any of the site’s sign-in credentials, a cookie or a header of any name, skips them:

Check skipped WCAG criterion
Content that appears on hover 1.4.13 Content on Hover or Focus
Single-key shortcuts 2.1.4 Character Key Shortcuts
Actions on pointer down 2.5.2 Pointer Cancellation

Everything else runs as usual. A person can still test those three on the signed-in pages and record the result in Review.

If your sign-in is a multi-step form rather than something a cookie can stand in for, a journey (Pro and Enterprise) can walk it: fill the email and password fields, submit, and carry on through the flow, with a scan at each step.

A value typed by a journey step is saved with the journey’s steps and shown as plain text to anyone who opens that step. It is not encrypted like the site’s credentials. Use the encrypted credentials wherever a cookie or header will do, and if a journey has to type a password, make it a test account’s.

  • Use a test account, never a real person’s. Give it the least access that shows the pages you need, and no admin rights. The scanner loads pages and, if you set them up, runs your interaction scenarios and journeys; anything those can trigger, the account can do.
  • Prefer a staging copy of the site if it has the same pages.
  • Expect the session to expire. When it does, scans quietly see the signed-out page. Save a fresh cookie, or use a long-lived token for the test account.
  • Mind what is on screen. Screenshots and the HTML of failing elements are kept with the findings, so the test account should hold test data, not customers’ details.
  • Remove access you no longer need: Clear the credentials, and revoke the token or end the session on your side.