XML Sitemap Validator

What gets checked

A valid sitemap is well-formed XML with the correct namespace declaration on the root element. Missing or misspelled namespaces are the most frequent cause of a sitemap being rejected outright, and the error message in Search Console is often unhelpfully generic. Each url element needs exactly one loc, and every URL must be absolute, including the scheme and host.

Escaping is the silent killer

URLs inside loc must be XML-escaped. An ampersand in a query string has to be written as an entity, and a raw ampersand makes the entire document invalid XML, so the whole sitemap fails rather than the one URL. The same applies to angle brackets and quotes. Any sitemap generated by string concatenation rather than a real XML writer is likely to have this bug the moment a URL contains a parameter.

Same host, same protocol

A sitemap may only list URLs on the same host as the sitemap itself, unless cross-submission has been set up through verified ownership of both. Mixing http and https, or listing www URLs in a sitemap served from the bare domain, causes those entries to be ignored. Trailing-slash inconsistency between the sitemap and the page's own canonical is a subtler version of the same problem.

Limits

Fifty thousand URLs and 50 MB uncompressed per file. Gzipping is supported and the 50 MB limit applies to the uncompressed size. A sitemap index may itself reference 50,000 sitemaps, and index files cannot be nested inside other index files.

Date format

The lastmod value must be W3C Datetime: either a complete date as YYYY-MM-DD, or a full timestamp with a timezone offset. Common failures are using a local format such as DD/MM/YYYY, or emitting a timestamp with no offset, both of which are rejected.

Content problems a validator cannot see

A structurally perfect sitemap can still be wrong. Listed URLs that 404, redirect, carry a noindex tag or canonicalise elsewhere are all valid XML and all reported as errors by Search Console. After the structure passes, crawl the listed URLs and confirm each returns 200 and is genuinely indexable.