# Platform and spec checks

Beyond “is this page working”, Monitor checks how each page conforms to the conventions search engines, crawlers and assistants actually rely on.

## What is checked

The checks run against the HTML Monitor already fetched, so they cost nothing extra and every page gets all of them on every crawl.

**Indexability.** Robots directives, canonical, whether the page is reachable from the sitemap, whether `robots.txt` permits it.

**Metadata.** Title and description present, unique across the site, and a sensible length. Duplicates across pages are reported as duplicates rather than as a problem on each page in isolation.

**Structure.** One `h1`, headings that descend without skipping, landmarks present.

**Structured data.** Schema found on the page, whether it parses, and whether required properties for its type are there.

**Social.** Open Graph and Twitter/X tags, including whether the image referenced actually exists — a broken OG image is invisible until someone shares the page.

**Links.** Internal links resolving, external links reachable, redirect chains.

**Rendering.** The rendered DOM compared against the raw source, which is how you find a page that looks fine in a browser and is empty to anything that does not run JavaScript.

## Why conformance rather than a score

A single number tells you a site got worse without telling you what to do. These are individual, named checks with a page and an expected value, so each one either passes or gives you something specific to change.

## Reading the results

Per page, or across the site to spot the systematic ones. A check failing on one page is an editing mistake; the same check failing on every page is a template, and one fix clears the lot.

## Related

- [Issues and severity](/docs/issues-and-severity)
- [Pages and crawl activity](/docs/pages-and-crawl-activity)
- [Changes and diffs](/docs/changes-and-diffs)