I built a website accessibility scanner that explains every fix in plain English

Most accessibility scanners tell you what is wrong in a language only a specialist reads: color-contrast, label, aria-required-children, plus a link to a WCAG criterion. That's fine if you live in this world. It isn't much help if you run a small business site, or you build sites for clients and need to tell them what to change.

So I built AccessLens: you enter a URL, and it tells you what's wrong, how much it matters, and how to fix it, in plain English.

What it does

• Opens your site in a real browser (Playwright) and runs axe-core against WCAG 2.1
• Gives the site a score out of 100
• Ranks the problems by how much they matter, so you know what to fix first
• Explains each problem in plain language instead of a rule ID
• Detects Shopify, WordPress, Wix and Squarespace sites and gives step-by-step fix instructions for those platforms

How it's built

• Next.js (App Router, TypeScript) and Tailwind
• PostgreSQL with Drizzle ORM
• Playwright and axe-core for the scanning
• No Redis or job queue. A scan runs as a background task after the response is sent, and the frontend polls for the result. It's the simplest thing that works at low volume.
• It runs on a small VPS, because Playwright plus a real Chromium doesn't fit serverless very well

What it can't do

Automated scanning only catches part of the accessibility problems on a site. It can't tell whether your alt text is actually meaningful, or whether the keyboard experience makes sense. Think of a scan as a starting point, not a compliance certificate.

Try it

It's in early access and free to try: scanner.analytihq.com

I'm Roshan, a .NET and RPA developer from Sri Lanka, and this is my first product aimed at web teams. I'd love your feedback:

• What would you want a scanner like this to check that the usual ones miss?
• If you build sites for clients, what would make a report ready to send to a client?