Guide baseline: v0.9.11
Web Scanner
Run automated accessibility checks against rendered pages and use the findings to guide remediation and manual testing.
Automated coverage
Playwright renders JavaScript pages and axe-core supplies automated checks. Results depend on the engine, page state, content reached and selected scan mode. Automated findings do not certify WCAG conformance or exercise every authenticated route, interaction and dynamic state.
Crawling and proposed code
Use max_depth and max_pages to bound the crawl. mode selects the scan mode; the example disables generated code fixes so that it does not require that AI lane. Where enabled and configured, suggested HTML/CSS changes need developer review and testing in the application. Scanning does not deploy changes to the site.
Manual testing remains necessary
Check complete tasks with the keyboard and intended assistive technology, including focus order, dialogs, forms, error recovery and changing content. Inspect whether text alternatives convey meaning. A source-code check or suggested fix cannot substitute for testing the rendered result and the states users encounter.
Authenticated API usage
Use your deployment URL and an authorized API key. The localhost URL below is for a locally running API.
curl --fail --silent --show-error \
-H "Authorization: Bearer $AELIRA_API_KEY" \
-H "Content-Type: application/json" \
--data '{"url":"https://example.com","mode":"quick","max_depth":1,"max_pages":1,"generate_code_fixes":false,"capture_screenshots":true}' \
http://localhost:8000/education/web/scanThe POST returns a scan_id, not completed findings. Poll GET /education/scans/{scan_id}/progress, then GET /education/scans/{scan_id}. Handle failed work, timeouts and unavailable scores explicitly. The request model also exposes scan_images, scan_multimedia, scan_math and validate_alt_text; configure and authorize relevant providers before enabling AI options.
Evidence and review
This guide describes released Core source; availability in a hosted deployment depends on its release and configuration. Review output before publishing it.