A sweep of every document in the repository for links that point nowhere: internal links to documents that no longer exist, and optionally external URLs that no longer respond.
Click Scan for Broken Links. It confirms first, since it reads every document in the repository. Results are saved against the project, so the tab still shows the last scan when you come back to it, with the timestamp in the top right.
Check external URLs is off by default. Turning it on requests every outbound URL found in your content, which makes the scan considerably slower but catches dead third party pages. A new scan always replaces the previous results.
Links to Prismic documents that are missing or unpublished. Each row shows the document type, the UID of the document containing the link, its language, and the field path the link sits in. The UID links straight to that document in the Prismic page builder, so you can open it and fix the field.
External URLs that did not respond successfully, grouped by URL rather than by document. Each group shows the HTTP status returned and then every document using that URL, so one dead link shared across fifty pages reads as one row with fifty references rather than fifty rows.
Treat the status as a starting point, not a verdict. Sites that block automated requests can report 403 or 429 while working fine in a browser, so confirm before editing content.
Export CSV and Export XML download the current results, internal and external together. Both include the document type, UID, document ID, language, field path, the external URL and status where applicable, and a direct Prismic page builder link per row - which makes the CSV usable as a work list for someone editing content.
Only the Prismic repository name. No Vercel credentials are involved, so this tab works before the rest of the project is configured.