XML Folio

XML Folio / Notebook

A reading copy is not your sitemap.

Keep the machine URL. Publish a separate view when people need to inspect it.

XML Folio editorial notes · Reviewed October 15, 2026

Preserve the job of the original file

A sitemap communicates URLs and associated metadata to consumers such as search crawlers. The HTML reading copy has a different purpose: it makes selected values easier for a person to review. Do not submit the generated HTML in place of an XML sitemap or replace the existing endpoint with an HTML response.

The Sitemaps protocol defines a namespaced urlset containing url entries with loc values. XML Folio requires that namespace so an unrelated element named “url” is not silently interpreted as a sitemap. It also supports a namespaced sitemapindex, which lists child sitemap URLs.

Work through the two-URL fixture

Load “Sitemap · a styled URL list” in the workbench or download the sample XML. Its first URL contains a query string with an escaped ampersand. The exported anchor contains the logical & separator once; it does not turn it into an extra encoded text fragment. The second URL contains space%20notes, which remains URL-encoded.

The two lastmod values stay as source text. XML Folio does not replace them with the export date. The snapshot generation timestamp appears separately. This matters because a reading-page build is not evidence that each destination page was modified.

Inspect extensions before claiming fidelity

The sample’s second entry includes an image extension. The reading profile does not display that image or preserve its meaning. Its namespace-qualified element paths appear in the omission list. This is intentional evidence of a limitation, not a warning that the original sitemap is invalid.

For urlset, the mapped values are loc, lastmod, changefreq and priority. For sitemapindex, they are loc and lastmod. Images, video, news, alternate-language annotations and other extensions require their own review. Nested omitted elements are counted separately, so an omission count is not a count of lost pages.

Do not turn the preview into an audit claim

This tool does not crawl destinations, test HTTP status, validate date syntax, compare hosts, infer canonical URLs or inspect index coverage. An HTTP(S) link may be syntactically usable and still return an error. A row appearing in the reading copy proves only that the selected source field was mapped.

The local tool accepts at most 1 MiB and 1,000 records; these are workbench limits, not the protocol’s own maximums. Large collections should be split or processed in a build pipeline. A sitemap index export shows child maps without fetching or recursively merging them.

A practical release sequence

  1. Save the existing XML response and its content type.
  2. Generate a reading copy and review every omission category that matters to your audience.
  3. Host the HTML at a distinct human URL, keeping the XML response untouched.
  4. Compare the expected links in a browser, then test the XML with its real consumer.
  5. Regenerate the static page when the source changes, or choose the optional same-origin refresh behavior.

Conversion alone does not improve crawling, indexing or rankings. Read the publication guide for a separate-browser check and clear freshness labeling.