Overview

Equall exposes an MCP server so your AI coding agent (Cursor, Claude, …) can drive the whole accessibility loop from your editor — without ever leaving your workflow.

Once connected, your agent has six tools and can run the full loop on its own, so you rarely need to open the dashboard:

• Connect a repository — add a GitHub repo as an Equall project.
• Run a scan — scan the committed code and update issue statuses.
• Conformance summary — read the per-criterion WCAG audit: what passes, fails, or needs human testing.
• List issues — read the accessibility findings back, tracked across scans.
• List pages — read the site's page registry (routes, sitemap URLs, manual entries) and their legal classification.
• Add a page — register a page the automatic detection missed.

The loop, at a glance

A typical session, all driven by the agent:

1. Connect the repo (skip if it's already an Equall project):

   "Connect acme/dashboard to Equall."

2. Scan it:

   "Run an accessibility scan on acme/dashboard."

3. Wait for the scan to finish, then read the audit and the issues:

   "Which WCAG criteria are blocking acme/dashboard, and list the issues for the worst one."

After you push fixes, just repeat steps 2–3 and ask the agent what changed.

Next

Start with Set up the connection to point your agent at the server.