How to keep your Knowledge Base current

By the end of this you will have looked at every source that needs attention, refreshed the ones that changed, and retired the ones that no longer apply. A monthly pass over thirty sources takes about fifteen minutes. None of these actions spends article credits or changes articles you have already written.

Before you start

• Know the labels. Each row shows a usage label and a status, such as Brand knowledge · Facts allowed · Ready · refresh failed. The statuses that need you are Needs attention, Review needed, Ready · refresh failed, Legacy index · refresh needed, Legacy index, and Removed.
• Have the newer file to hand if you plan to replace a document.
• Decide what stale means for you. A pricing page belongs on Daily or Weekly; a research report can wait for its Review by date.

Steps

1. 1

   Scan for warnings

   Open Knowledge Base from the sidebar. Amber lines under a row are warnings, and any status other than Ready deserves a look. Use the Documents, Added URLs, and Website filters to take one kind at a time.

2. 2

   Re-check a URL

   Open the source and press Check URL now. A strip in the drawer shows progress. If the page changed, the row reads Ready · updating until the new version activates. If not, Last checked in the detail grid updates and the version number stays the same.

3. 3

   Reprocess without re-fetching

   Reprocess saved content, shown on URLs with a verified snapshot, re-indexes the saved full-page text without contacting the site. Use it after an indexing failure when the page itself is fine. For a document, Reprocess original does the same with the stored file.

4. 4

   Replace a document

   Press Replace document and pick the new file. The drawer shows Uploading replacement with a percentage, then Upload received; queueing replacement…. The active version stays available until the replacement is ready; then Version history shows the new number marked Active.

5. 5

   Extend or clear a review date

   For a row reading Review needed, read the Preview and confirm the content is still true. Expand Source settings, set a new Review by or clear it, and press Save settings. The status returns to Ready at once. Checking a URL never extends its review date; only this edit does.

6. 6

   Pause or resume

   Press Pause use to keep a source but stop it informing writing. The status reads Paused, the button becomes Resume use, and scheduled checks stop. Press Resume use to bring it back.

7. 7

   Delete

   Press Delete. Serpio asks Remove this source from future article writing? Existing articles and idea collections will stay unchanged. Confirm, and the row disappears; its versions, passages, and stored files are cleaned up. There is no undo.

8. 8

   Refresh the website crawler

   Expand Website crawler at the top of the page and press Refresh. The panel reads Crawling your site… for 30 to 90 seconds, then shows the new pages indexed count and last refreshed time. Only pages whose content changed get a new version.

9. 9

   Read Version history

   In any drawer, Version history lists versions newest first with their state: queued, processing, ready, failed, or superseded. The one marked Active is what retrieval uses. A superseded entry right after a check means the content had not changed; a failed entry shows its error beneath it.

What happens next

Every action above that changes what is active bumps the brand's knowledge revision. Articles started from now on retrieve only the current versions; idea collections and existing articles, including their Knowledge used lists and quality reports, stay as they were. Between passes Serpio keeps working: Daily and Weekly URLs are checked on schedule, the site is recrawled every Sunday, and interrupted processing is retried up to three times.

If something looks off

• Ready · refresh failed keeps returning on a URL. Read Latest processing issue in the drawer. Common causes are the page now returning an error, exceeding the 4 MiB fetch limit, or no longer serving HTML. The last good version still serves meanwhile.
• A source turned Removed. The page answered 404 or 410. If it moved, add the new address as a URL source and delete the old one.
• Reprocess saved content is not shown. There is no verified snapshot yet, or the page is Removed. Press Check URL now.
• The version number did not change after a check. The content was identical and the version was marked superseded, which is expected.
• The site index count went down after Refresh. Discovery is capped at about 100 pages and prioritises product and support pages. Add dropped pages individually under Add URL.

Go deeper

• Knowledge Base overview for how sources inform articles.
• Adding documents for formats and the replacement flow.
• Usage settings and permissions for what to check when settings are pending.
• Versions, statuses, and source actions for the full status table and detail grid.
• Serpio's site index is not Google's index for what the crawler indexes and why pages go missing.