Skip to main content

Run the automated scanner on one sample or on every sample of an audit, and what it writes.

Goal

Scan one sample, or all samples of an audit, and turn the results into findings.

Where the scanner is

Inside an audit opened from My Audits, never on the Dashboard or a settings page. The two scan buttons show only to members who can edit the audit, only while it is open (not closed), and only on WCAG, RGAA and RAWeb audits: RAAM (mobile) and RAPDF (PDF) samples are not web pages. Without the scanner on the workspace, the buttons carry a lock icon and open a short presentation of the Scanner add-on.

Scan one sample (with review)

  1. Open the audit from My Audits, then the sample.
  2. In the sample header, click Scan (French: « Scanner »).
  3. Wait (usually seconds). The button shows the issue count and a View results button (« Voir les résultats »).
  4. The Scan results panel (« Résultats du scan ») groups results by criterion:
    • Detected failures (« Échecs détectés ») are pre-selected.
    • Needs review (« À vérifier ») items are unselected: tick those to flag as Needs review; the rest stay for manual evaluation.
  5. Click Apply N criteria (« Appliquer N critères »). Detected failures become Failure, ticked items Needs review, each with a problem text citing the automated scan.

Each result lists up to 20 affected elements per rule (CSS selector, element code, what fails). If the audit has no technology set and the page looks built with one (React, Vue, Angular, WordPress…), the panel offers it for the audit's code fixes.

Applying also records the elements on each criterion and, when the rule has a known fix, writes the criterion's Expert proposal in the audit's technology. See product / findings.

Scan all samples (no review step)

  1. Open the audit from My Audits.
  2. In the audit header, click Scan all samples (« Analyser tous les échantillons »).
  3. Choose samples and start. They are scanned one after the other.

Results apply directly: detected failures become Failure, criteria with only uncertain results become Needs review. The batch stops after 2 failed scans in a row (usually rate limit or connection); failed samples can be retried from the summary.

What a scan may overwrite

Only criteria still Not tested, or filled by an earlier scan and not commented on since. Criteria an auditor evaluated are left as they are.

Verify success

  • Applied criteria show their new status and the scanner's problem text.
  • Criteria evaluated by hand are unchanged.

Common pitfalls

  • Trusting the scanner blindly. Automated testing catches roughly 30 to 50 percent of WCAG failures; the rest is manual. Ambiguous results belong in Needs review, not Success. After Scan all samples, read each sample's applied findings: nobody reviewed them before they were written.
  • Leaving scanner-generated text unedited. It often needs rewriting for clarity. See methodology / finding quality.
  • Scanning behind authentication without the workspace's own Browserless key configured for it. The shared scanner only sees public pages. Set the key in My Workspaces > Integrations.
  • Looking for a scan button on a RAAM or RAPDF audit. There is none, by design.

Last reviewed: 2026-11-02