AccessPatch – Accessibility code audit inside AI dev tools

Embed one-click WCAG checks in AI coding assistants so generated frontend code is compliant

Problem

Developers using AI tools to generate frontend code often overlook accessibility rules, creating unfriendly sites and compliance risks. Existing accessibility audit tools require manual runs or CI integration, and AI coding assistants have no built-in checks.

Solution

An MCP server that lets AI dev tools like Claude Desktop and Cursor directly call a local WCAG audit engine. When AI generates HTML/CSS/JS, it automatically checks contrast, semantic tags, ARIA attributes, etc., and returns fix suggestions or directly modifies code.

Users

Frontend developers using AI coding assistants, accessibility-conscious startup teams, government/education contractors

MVP

Implement a local MCP server wrapping axe-core or similar; support scanning HTML snippets and local pages; return issue lists with position markers; provide text fix suggestions; integration docs for Cursor and Claude Desktop.

Revenue

Open-source core engine + paid premium (team rule sets, CI integration, report export); or $9/month per developer subscription.

Why now

AI-generated code is becoming mainstream, yet accessibility compliance (e.g., European Accessibility Act) is tightening. Shifting checks left in the coding phase costs far less than remediation, and the MCP ecosystem is nascent—indie developers can claim space quickly.