Official foundations
Standards bodies, official documentation and public authorities are used as primary reference points.
Quality tools reveal speed, accessibility and stability issues, but no score replaces a real person completing a task. Compare lab and field data when the latter is available.
Measure one page on a phone, then complete a task by keyboard and with larger text. Link findings to a specific component or piece of content before fixing them.
AI-generated illustration
Selection prioritises recognition within the field, professional use and reference status. Highly specialised documentation entries are consolidated under their main resource.
Standards bodies, official documentation and public authorities are used as primary reference points.
Established services are grouped by actual use: building, hosting, measuring, securing, publishing or promoting a website.
Inclusion is editorial and unpaid. A listing does not imply a partnership or guarantee.
| Criterion | Question to resolve |
|---|---|
| Test type | Does the result come from a laboratory test, field data or human assessment? |
| Journey | Do contact, search and navigation remain usable with a keyboard and on mobile? |
| Priority | Which problem blocks a user and which fix can be verified? |
Keep a short decision record: requirement, source, observed result and unresolved point. Test the same task with each candidate instead of comparing marketing lists of features.
Compare their role and documentation before selecting a tool. Inclusion is not a ranking.
11 resources displayed
The date applies to the linked document, not to a review of every service feature.
The link restores the search and filters, including resources selected for comparison.
No resource matches these filters. Try a broader search.
Select resources from the cards below. Compare their role and practical use, then verify current terms, costs, data export and support on official websites. Geographical origin describes the publisher; it does not establish where your data will be hosted.
No resource selected.
Learning paths covering Web development fundamentals, performance and accessibility. Useful for studying a topic through explained examples. Select a module for the task to improve, then compare the exercise with your own site’s content, constraints and users.
web.dev · Source consulted
Use it for : Connect performance or accessibility guidance to a measurement on your pages.
Documentation for the Chrome browser, DevTools and Web platform.
Chrome for Developers · Source consulted
Use it for : Reproduce an issue in the browser and record the test steps.
URL analysis combining Lighthouse diagnostics with field data where available. Useful for distinguishing a simulated trial from historical observations at page or origin level. Record the displayed scope and test conditions before comparing results.
PageSpeed Insights · Source consulted
Use it for : Separate field and lab data, then check the relevant URL and device before prioritising fixes.
Learning example: adapt files, identifiers and conditions to a test copy before interpreting the result.
Automated page checks for performance, accessibility, best practices and SEO. Useful for identifying reproducible issues and tracking fixes under defined conditions. Scores replace neither journey testing nor a full accessibility evaluation and can vary between runs.
Lighthouse · Source consulted
Use it for : Use diagnostics to find reproducible causes and compare tests under similar conditions.
Learning example: adapt files, identifiers and conditions to a test copy before interpreting the result.
Official documentation · Chrome — Lighthouse performance scoring
Detailed page-load testing from multiple devices and locations.
WebPageTest — GitHub · Source consulted
Use it for : Inspect loading and network waterfall for a defined scenario and location.
Verification path
Learning example: adapt files, identifiers and conditions to a test copy before interpreting the result.
| Run | First visit | Repeat visit |
|---|---|---|
| 1 | 3.2 | 1.8 |
| 2 | 3.4 | 2.0 |
| 3 | 3.3 | 1.9 |
Checks document markup syntax and explains detected errors. A valid result is neither a quality seal nor a complete accessibility assessment.
W3C — Markup Validator Help · Source consulted
Use it for : Start with the first errors, which may produce cascading messages. Check that the tested document matches the served page, then revisit keyboard use, forms and reading: these journeys require checks beyond syntax.
Learning example: adapt files, identifiers and conditions to a test copy before interpreting the result.
POST request body can contain the HTML directly, with Content-Type: text/html; charset=utf-8. The out=json parameter requests a structured report. Distinguish the submitted document from the returned report and retain the examined page version.messages array and the type field. An error entry identifies a detected problem; non-document-error means checking could not complete. The latter must not be presented as an error-free page: the result remains indeterminate.Browser- and version-specific support tables for Web technologies. Useful before choosing an API or CSS feature and for reading compatibility qualifications. Usage shares depend on the selected population; then test the feature and fallback in required environments.
Can I Use · Source consulted
Use it for : Check feature support for the browsers used by the intended audience.
Verification path
Learning example: adapt files, identifiers and conditions to a test copy before interpreting the result.
subgrid. Record the exact feature and target browsers and versions before reading the table. A global percentage alone does not describe the project’s audience.Website and application testing on real browsers and devices.
BrowserStack · Source consulted
Use it for : Replay an important journey on priority devices and browsers.
Deque tools for automating accessibility checks in different development and testing contexts. Useful for finding known issues in the states examined. Identify the product, scope and remaining manual checks instead of inferring general conformance from no alerts.
axe · Source consulted
Use it for : Pair automated findings with human review of the journey.
Verification path
Learning example: adapt files, identifiers and conditions to a test copy before interpreting the result.
axe-core report contains violations: [] and an incomplete result. The latter needs further checking; it does not become a pass because the violations array is empty.inapplicable means no content matching that rule was found. Complement automation with keyboard journeys and relevant human checks.Visual page evaluation highlighting accessibility findings and information needing manual review. Useful for locating elements, structure and alternative-text meaning. No automated tool checks every criterion; supplement it with reading, keyboard use and actual tasks.
WAVE · Source consulted
Use it for : Review alerts in context, then add keyboard, screen-reader and journey-comprehension checks.
Learning example: adapt files, identifiers and conditions to a test copy before interpreting the result.
Resources, training and tools dedicated to Web accessibility.
WebAIM · Source consulted
Use it for : Read an accessibility method, then apply it to a real page.
These examples illustrate uses, not rankings or endorsements. Verify each provider’s current documentation and terms before a decision.
Quality tools reveal speed, accessibility and stability issues, but no score replaces a real person completing a task. Compare lab and field data when the latter is available.
Questions to checkA practical use : Measure one page on a phone, then complete a task by keyboard and with larger text. Link findings to a specific component or piece of content before fixing them.
What this cannot prove : A lab score varies with its environment and cannot describe all visitors. Automated checks do not cover every accessibility barrier.
Examples : web.dev · PageSpeed Insights
Related guide : Test website accessibility through real tasks
Quality tools help locate a defect; validation concerns a task that can actually be completed. Preserve the device, width, context and actions for each trial. These situations connect a score or file to an observable difficulty.
Reference : W3C · Reflow
Before making a resource part of website operations, plan for updates and an exit too. These checks concern your trial results without assuming features or prices of a particular plan.
Retain URL, viewport, browser, date and measurement mode. A lab result without conditions cannot be compared reliably with another run.
For each retained defect, keep steps, expected result and impact. A score disappearing from a tool does not prove a fix.
Repeat representative manual checks. A tool update can change its rules even when the page is unchanged.
This non-exhaustive directory had its data consistency and links checked on 2 October 2026. Some websites restrict automated checks. It contains external links to independent websites. Their availability, content, prices, cookies and terms may change. Française du Numérique does not control them and receives no commission for their inclusion.
Before choosing a resource, run a limited trial using the same data and task. Keep the outcome and its limitations rather than a ranking based on familiarity.
Run the same task on mobile, with a keyboard and using the chosen inspection tools. Keep page, device, conditions and observed obstacle. An automated score does not replace a journey check; connect each issue to a tested correction.
An image’s meaning and the version actually displayed concern different aspects of quality.
| Reference | Value or syntax | Scope |
|---|---|---|
| WCAG 2.2 · 1.4.3 | 4.5:1 / 3:1 | Normal / large text (18 pt or 14 pt bold); criterion exceptions apply. |
| WCAG 2.2 · 2.5.8 | 24 × 24 CSS px | A smaller target can satisfy the criterion through spacing; other exceptions apply. |
| WCAG 2.2 · 1.4.10 | 320 CSS px | Vertical content without two-dimensional scrolling; exceptions cover inherently two-dimensional content. |
| WCAG 2.2 · 3.3.7 | A · same process | Within one process, previously supplied information must be auto-populated or available to select, subject to the criterion’s exceptions. Browser autocomplete alone is insufficient; the site must offer the information already received or avoid requesting it again. |
Dates indicate when references were consulted. These facts are neither a site audit nor an overall compliance statement.