On-page SEO is everything on the page itself that helps search engines understand it and helps people decide to click: the title tag, the meta description, a clear H1 and sensible subheadings, enough useful text, descriptive links, image alt text, a correct canonical tag and no accidental noindex. When one of these is wrong, good content can sit on page five without anyone noticing why.
Enter a page address and, if you like, the main phrase the page should rank for. The analyzer reads the live page once through our server and checks the title and description length, the number of H1s and skipped heading levels, how much main content there is (menus, headers and footers are ignored), whether the canonical points to the page itself, whether the page or the server header says noindex for Google, images without alt text or without width and height, empty or vague link text, internal and external links, and any JSON-LD structured data, including syntax errors.
With a keyword, it also shows whether the phrase appears in the title, H1, first 100 words, meta description, URL and a subheading, and how often it is used. There is no ideal keyword density, so treat the count as a check for missing or overused wording, not a target. The tool cannot judge whether your content is the best answer for the searcher; that still needs a human read.