Type any page address to audit it the way a search engine reads it: status and redirects, canonical and robots rules, title and meta description, headings outline, word count, images and alt text, links, structured data and social sharing tags. Add a keyword to see where it's used. Every problem comes with what to do about it.
On-page SEO is everything on the page itself that helps search engines find, understand and show it: whether it can be indexed, its title and description, how its headings are organised, its text, images and links, and the structured data and sharing tags in its code. This tool fetches any public page once and checks all of it in one report, with a score for each area, the headings outline, every image and its alt text, and every link. Add a keyword to see where the page uses it.
Tip: Start with Indexing. A page that's marked noindex, answers an error or has a canonical pointing elsewhere won't rank however good the rest is, so those few checks matter more than everything else in the report.
Each check is worth one point when it passes, half when it's a warning and none when it's a problem; the score is the share of points, out of 100, for each area and for the whole page. It's a quick way to compare pages and track progress, not a prediction of rankings. A page can score 100 and still not rank if the content doesn't answer the search better than other pages, and a page with a few warnings can rank first.
Treat red items as things to fix and amber items as things to consider. Some warnings are fine for a particular page: a contact page doesn't need 300 words, and a canonical pointing elsewhere is correct on a duplicate.
The checker reads the HTML the server sends, like most crawlers do on first fetch. If your site builds content with JavaScript in the browser, text, links or tags added that way won't be counted here; Google usually renders JavaScript, but later and not always fully, so important content and tags are safest in the server's HTML. Speed and Core Web Vitals depend on the visitor's device and network, so measure them with PageSpeed Insights or the Core Web Vitals report in Search Console. Pages behind a login, or that block automated requests, can't be checked.
No or duplicate title: give every page a unique title that says what it offers, under about 600 pixels. Missing description: write one or two sentences that make someone want to click. Several H1s: keep one main heading and turn the others into H2s. Missing alt text: describe what the image shows in a few words, or use alt="" for purely decorative images. Canonical problems: make each page's canonical its own final address, and point duplicates at the main version. Vague links: replace "click here" with words that describe the destination. No structured data: add the schema type that fits the page, such as Article, Product or LocalBusiness, with the schema generator.
Problems usually come from templates. Check one page of each type (home, category, product, article) to find site-wide issues.
Check the page that ranks first for your keyword and compare headings, length and schema.
Google indexes the mobile version. A missing viewport tag is a serious problem.
If the title, H1 and top phrases point at different topics, decide what the page is for.
Releases can remove canonicals or add noindex by mistake. A quick check catches it.
Open the schema validator, hreflang checker or broken link checker for any area that needs a closer look.
No, it checks one page at a time, in depth. Check your main templates to find most site-wide issues.
Navigation, header, footer, side panels and scripts are left out, so only the main content is counted.
The big heading may be styled text rather than an <h1>, or it may be added by JavaScript after the page loads.
No. It means the page has no technical or on-page issues the tool can see. Rankings depend on relevance, quality, links and competition.
No. Only public pages can be fetched.
No. The page is fetched once by our server and checked in your browser. Nothing is kept.
Updated