Log inSubmit server
A11Y Toolkit logo

A11Y Toolkit

Audits, fixes, documents, and monitors WCAG 2.2 accessibility through 37 MCP tools and a zero-dependency core.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install A11Y Toolkit

Pick your client, then add it in one click or paste the config.

STDIOPython
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "a11y-toolkit": {
      "command": "uvx",
      "args": [
        "--from",
        "a11y-toolkit",
        "a11y-toolkit-mcp"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Install not yet sandbox-tested— we haven't run this command through our automated sandbox yet. Not a red flag; we're working through the catalog.

Overview

A11Y Toolkit MCP server gives coding agents tools for WCAG 2.2 audits, remediation, evidence collection, accessibility documents, and regression monitoring. Its 37 tools cover static pages, rendered browser behavior, PDFs, DOCX files, contrast, keyboard interaction, forms, screen-reader output, and more. Core operations run without dependencies, while rendered checks use Playwright and a browser such as Chromium, Firefox, WebKit, Chrome, or Edge. Reach for it when an agent needs repeatable accessibility checks and human-reviewable evidence rather than a one-time scan.

Use cases

  • Audit a web page against WCAG 2.2
  • Test keyboard, form, reflow, and screen-reader behavior
  • Generate accessibility statements and VPAT reports
  • Compare accessibility snapshots between releases
  • Create a human-signable evidence pack

Key features

  • 37 MCP tools for WCAG 2.2 workflows
  • Zero-dependency static audits and evidence generation
  • Playwright-backed rendered interaction checks
  • Contrast analysis for images, gradients, and live elements
  • Accessibility documents, reports, and evidence exports
  • English and Spanish output support

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 1d ago

Recent health check succeeded.

9/9 checks healthy over the last 57d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 12d ago via OSV.dev · a11y-toolkit (PyPI)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars3Stargazers on the source repository.
Last commit2d agoMost recent push to the default branch.

How A11Y Toolkit works

What A11Y Toolkit does

A11Y Toolkit brings accessibility testing into an MCP-compatible coding workflow. It covers the WCAG 2.2 loop: inspect a page or document, identify issues, apply safe fixes, create conformance materials, and compare later results for regressions. The package exposes 37 MCP tools, six prompts, and a skill.

Audit coverage includes URL and DOM checks, PDF and DOCX screening, HTML validation, accessible-name lookup, and WCAG criterion explanations. Interaction checks examine keyboard traps, form errors, hover dismissal, 320px reflow, infinite scrolling, reduced motion, forced colors, captions, and screen-reader-oriented output. A blindfold journey tool evaluates tasks using only what a screen reader exposes.

How it works

Static operations run with the package's zero-dependency core. Browser-backed operations render the target page and can inspect behavior that static analysis cannot, including image or gradient contrast, focus movement, submissions, tooltips, and responsive layouts. Playwright is required for these rendered checks, with browser installation handled separately.

Findings can be followed by deterministic autofixes, contrast calculations, passing-color suggestions, or sealed patch sets keyed by hashes. The watch-oriented tools create accessibility snapshots, record tab order, compare snapshots, and compare staging with production URLs. Evidence tools produce a human-signable pack with hashes, a coverage matrix, and an empty signature block; other outputs include EARL exports, OpenACR YAML, VPAT 2.5 material, legal statements, and an SVG badge.

Results and generated content support English and Spanish through the lang argument where applicable. The toolkit also includes a WCAG 3 draft preview that reorganizes WCAG 2.2 findings without presenting a fabricated score, plus a seeded mutant bank for checking detection quality with kill-rate and false-positive measurements.

Setup and configuration

For MCP clients, the documented launch command uses uvx:

Terminal
claude mcp add a11y-toolkit -- uvx --from a11y-toolkit a11y-toolkit-mcp

The same engine is available as a command-line package. Install it with pip install a11y-toolkit, then run commands such as a11ytoolkit audit --url https://example.com or a11ytoolkit pair "#1f2328" "#fbfaf7".

Static audits, contrast calculations, PDF and DOCX screening, and evidence generation need no extra dependencies. Install Playwright and a browser for rendered checks:

Terminal
pip install playwright
playwright install chromium

Tools and capabilities

  • Audit URLs, rendered DOM, PDFs, and DOCX files.
  • Test contrast pairs, image backgrounds, live DOM elements, and suggested replacement colors.
  • Detect keyboard traps, form error behavior, hover-dismissal issues, reflow problems, and infinite-scroll failures.
  • Produce screen-reader transcripts, accessibility statements, VPAT 2.5 reports, OpenACR YAML, and evidence packs.
  • Track snapshots, tab order, staging-versus-production differences, and aria-live behavior.
  • Export evidence for ACT Rules interoperability and generate WCAG 3 draft groupings.

Limitations and notes

Automated signals cover 51 of the 55 WCAG 2.2 A and AA criteria. The project explicitly positions the machine as a finding mechanism and retains a human sign-off step for conformance artifacts. Browser-dependent checks require Playwright and a browser installation, unlike the zero-dependency core operations.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about A11Y Toolkit

A11Y Toolkit is an MCP server that connects AI coding agents to accessibility audits and WCAG 2.2 workflow tools. Its 37 tools cover page, DOM, PDF, and DOCX checks, contrast analysis, keyboard and form testing, screen-reader transcripts, remediation, evidence packs, documentation, and regression monitoring.

Install it in an MCP client with `claude mcp add a11y-toolkit -- uvx --from a11y-toolkit a11y-toolkit-mcp`. The package also provides a CLI installable with `pip install a11y-toolkit`. Add Playwright and a browser separately for rendered checks.

Yes, A11Y Toolkit is intended to work with Cursor and other MCP clients. The quickstart lists Cursor alongside Claude, Codex, and Windsurf, and uses the standard MCP launch command.

Only rendered browser checks need Playwright. Static audits, contrast math, PDF and DOCX screening, and evidence-pack generation run with the zero-dependency core; rendered checks require Playwright and an installed browser.

Yes, it can generate EAA and RD 1112/2018 accessibility statements, VPAT 2.5 reports, OpenACR YAML reports, SVG badges, and evidence packs designed for human sign-off.

Related MCP servers

All in Developer Tools All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewA11Y Toolkit AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/a11y-toolkit?style=directory)](https://allmcps.com/mcp/a11y-toolkit)
HTML Embed
<a href="https://allmcps.com/mcp/a11y-toolkit"><img src="https://allmcps.com/api/badge/a11y-toolkit?style=directory" alt="A11Y Toolkit on AllMCPs" /></a>