AccessiGuard – Web accessibility scanner with AI fix suggestions

Hey HN,

I quit my engineering manager job last week. Family of 4, no safety net. Built and launched AccessiGuard in 6 days — first commit to production.

It's a WCAG 2.1 scanner that checks websites for accessibility issues and gives you actual fix suggestions (code snippets, not just "fix this").

How it works: - Fetches your page, parses the HTML with Cheerio, runs 22 custom checks against WCAG 2.1 AA criteria - Crawls linked pages on the same domain for multi-page scans - AI fix suggestions via OpenAI — points you to the exact element and tells you what to change - PDF reports, score tracking over time, monitoring for regressions

What it catches: Missing alt text, form labels, heading structure, ARIA issues, viewport zoom blocking, skip nav, duplicate IDs, generic link text, table headers, and more. Full list: https://accessiguard.app/scan-coverage

What it doesn't catch (being honest): No color contrast checking yet — that needs computed styles which requires a real browser. Currently on the roadmap alongside Puppeteer integration. Same goes for keyboard trap detection and SPA rendering.

Stack: Next.js 15, Supabase, Cheerio, OpenAI, Stripe, Vercel.

Pricing: Free tier is 5 scans/month. Paid starts at $29/mo. Agency plan is $99/mo for 50 sites.

Why another scanner? Most tools are either enterprise-priced, overlays that don't actually fix anything (the FTC fined accessiBe for this), or black boxes that don't tell you what they test. I wanted something transparent, affordable, and useful for developers who actually want to fix their code.

Would love feedback on scan accuracy, things I'm missing, or if the reports are actually useful. Happy to answer architecture questions.