Tools / Site accessibility crawl

Site accessibility crawl Pro

Audits up to 15 pages from a start URL and reports issues by how widespread they are, so shared-template problems surface first.

MCP tool name: audit_site_accessibility. REST: POST /api/a11y-site. Pro only.

Ask for it in plain words

Pro. Crawl up to 15 same-origin pages from a start URL, audit each one's static HTML for WCAG 2.2 problems, and report issues by how many pages they affect so that shared-template problems come first. Returns a markdown report.

Inputs

* required

Example

curl -X POST https://designers.allthepossibles.com/api/a11y-site \
  -H "content-type: application/json" \
  -H "Authorization: Bearer dz_YOUR_KEY" \
  -d '{"url":"https://allthepossibles.com","maxPages":6}'

Example output

The report field, for the input above (structured fields come back alongside it):

# Accessibility crawl of allthepossibles.com

Audited 5 of 5 pages (static HTML only), starting at https://allthepossibles.com/.

## Issues across the site, most widespread first within severity

- **HIGH** 3.1.1 Language of Page: html element has no lang attribute; screen readers guess the pronunciation. On 1 of 5 pages, 1 times.
- **HIGH** 3.3.2 Labels or Instructions: input type=password has no label (a placeholder is not a label). On 1 of 5 pages, 1 times. e.g. <input>
- **MEDIUM** 1.3.1 Info and Relationships: heading level jumps from h2 to h4. On 4 of 5 pages, 4 times. e.g. <h4>
- **MEDIUM** 1.3.1 Info and Relationships: no main landmark; screen reader users cannot jump to the content. On 1 of 5 pages, 1 times.
- **LOW** 2.4.1 Bypass Blocks: no skip link at the top of a page with many links. On 2 of 5 pages, 2 times.

## Pages

| Page | High | Medium | Low |
|---|---|---|---|
| / | 0 | 1 | 1 |
| /billing | 2 | 1 | 0 |
| /privacy | 0 | 1 | 0 |
| /terms | 0 | 1 | 0 |
| /support | 0 | 1 | 1 |

Issues that appear on most pages usually live in a shared template or component; fix those first. This crawl reads markup only and follows same-origin links from the start page, so it is a sample, not a complete inventory. It cannot test contrast, focus order, or script-built interfaces.

Limits of this tool

Deterministic: the same input always gives the same answer. It does not call a language model.