Screen reader checker
Screen readers and AI agents that browse do not look at your design. They read the accessibility tree: headings, landmarks and the names of links, buttons, images and form fields. See exactly what they get, and what they miss.
Or ask your AI assistant
This check is the screen_reader_view tool in the free OnPage.dev MCP server. Connect it to Claude, ChatGPT or Cursor once, then ask in your own words.
Read example.com/pricing the way a screen reader does and fix every link and button without a name.
screen_reader_view3 links announced without a namescreen_reader_view“Read more” 12 times, no main landmark
Why this matters for SEO
Accessibility is not a direct Google ranking factor. But the same things a screen reader needs are what search engines and AI read: alt text for images, link text that says where a link goes, real headings, and text that is not hidden. AI agents that shop and compare for people increasingly read pages through this same accessibility tree.
Well-built sites are not safe by default, including sites built with AI coding assistants. In a real scan of a carefully built news site that scored well everywhere else, the screen reader view still caught links announced as just “link” in nine different card layouts across the site: each thumbnail image linked to its article, but gave a screen reader nothing to say.
What gets flagged
- Links and buttons that are announced without a name, like an icon link or a thumbnail image without alt text.
- Vague link text such as “read more” or “click here”, which is also weak anchor text.
- Form fields without a label, a missing main landmark, no H1 and skipped heading levels.
- Visible text that screen readers do not get, often hidden by
aria-hidden.
Questions
Is this a full accessibility audit?
No. It shows what a screen reader announces and flags the most common problems. A full WCAG audit also covers contrast, keyboard use and more; the main OnPage.dev scan checks some of those too.
How is this different from an alt text checker?
It reads the page as assistive tech does after the page is rendered, so it also catches links and buttons that end up without a name, text hidden from readers and the real reading order.
More free tools
test_robotsllms.txt generatorGenerate an llms.txt for your website from its sitemap and home page in seconds.generate_llms_txtGoogle snippet checkerCheck if your title tag and meta description fit in Google on desktop and mobile, measured in pixels, and see where they get cut off.check_snippetRedirect checkerCheck status codes and full redirect chains for up to 20 URLs at once, and verify each one ends at the right page.check_urlsSchema markup validatorValidate JSON-LD structured data before you publish: syntax, @context and @type, ISO dates, absolute URLs and Google’s required fields for rich results.validate_schemallms.txt validatorCheck your llms.txt against the llmstxt.org format: title, summary, sections, link format, size and whether the linked pages load.validate_llms_txtEntity graph checkerSee how your structured data tells AI and Google who is behind a page: Organization, author, publisher and sameAs profile links, with the fixes as JSON-LD.check_entity_graphRedirect map generatorPaste your old and new URLs and get a 301 redirect map with a confidence score per match, as _redirects, nginx, Apache or Next.js rules.plan_redirectsMobile first-screen checkerLoad a page in a real browser on a phone, tablet and desktop and see what visitors get first: H1, call to action, pop-ups, tap targets, layout shift and what crawlers without JavaScript miss.render_pageSEO audit to Google SheetsScan a page and get every SEO finding as a Google Sheet with one formula, plus a ready Slack message and one task per issue for Notion, Linear, Jira or Asana.export_findings