Free URL Heading Extractor - Check H1 to H6 Headings of Any Web Page

What it does: a visitor pastes the address of a web page and sees every H1 to H6 heading of the page at once: as an outline tree, where each heading hangs under the heading it belongs to, so a broken structure stands out, as a table with the level, the text, its length, where it sits on the page (header, navigation, content, sidebar, footer), its #anchor, and its issues, and as a checklist of what to fix. The checks follow what Google and the W3C actually say: a page with no H1 or with empty headings gets a warning, several H1s, skipped levels (H2 to H4), long headings, and repeated headings get a tip, the H1 is shown next to the title tag. Headings the page hides (a mobile menu copy, aria-hidden) are marked and kept out of the checks, so they never count as a second H1. The visitor copies the outline, the checks, or a Markdown outline, or downloads the table as CSV.

Who it is for: SEO specialists auditing on-page structure, content writers and editors checking an article's outline before or after publishing, marketers comparing competitor pages, developers and accessibility testers checking heading order for screen readers. And site owners who want their own free heading checker page.

What you get: the complete tool as one Cloudflare Worker for your own account and domain: the page, a JSON API with the table, checks, and counts, and a plain-text API that answers the outline tree, with no database, no API key, no dependencies, the Cloudflare free plan is enough. Settings (name, color, sub-path, limits) are in one config file, every text and the check limits are in one file. Tests and HTML documentation included.

What you do: run npx wrangler deploy, or try it first with npx wrangler dev. Then share the page (links with ?url= run a check at once) or call api/run?url=...&format=text from your scripts.

Features

• Every h1 to h6 and every role="heading" element of the page in page order, hidden menus included
• Outline tree with tree lines, a Markdown outline, and a table: #, Level, Heading, Characters, Where, Anchor, Issues
• Checks without a made-up score: no H1 and empty headings (warnings), several H1s, skipped levels, long headings over 70 (H1) or 100 characters, repeated headings (tips), H1 next to the title tag
• Hidden headings (hidden attribute, aria-hidden, inline display:none) marked and left out of the checks, or left out of the list with one switch
• Clean heading text: permalink signs (¶, #, §) and edit links removed, image headings read from alt text, screen-reader-only and ARIA headings marked
• Whole page or main content only (the article's headings without menus, sidebars, and footers)
• Counts: headings, H1 to H6, warnings, tips, hidden
• Copy or download the outline (.txt), the checks (.txt), and the Markdown (.md), download the table as CSV (formula-safe for spreadsheets)
• Pages up to 5 MB and 5,000 headings, plain messages for pages built with JavaScript, PDFs, missing pages, sites that refuse tools, slow sites, and too many requests
• Safe by design: one page per request, no crawling, public addresses only (checked on every redirect), an honest User-Agent, a per-visitor limit, nothing stored
• Shareable links with ?url=, works on phones, with the keyboard, and inside a frame

Screenshots

Try it live: https://codemarket-live.codemarket.workers.dev/url-heading-extractor/ (no login needed).