Skip to main content
Auto-sync re-imports your website links on a schedule, so the chatbot’s answers follow your content without manual retraining.

​
What auto-sync covers

Two kinds of schedules keep Website Links content fresh:
  • Whole-import schedules re-run a website crawl or sitemap import, refreshing every page it produced.
  • Per-link schedules re-fetch an individual link. Set them by opening a link in Links List, or select several and use Update Config.
YouTube videos have no schedule; resync them manually from Links List (or sitegpt knowledge documents resync) when a video’s captions change. Uploaded files and connected apps (Notion, Google Drive, Dropbox, OneDrive, Box) do not run on a schedule. Resync them manually from Files & Data Sources when the source changes, or script it with the CLI: sitegpt knowledge documents resync. The exception is GitHub, which has its own auto-sync schedule and also discovers newly added files. Text snippets have no schedule either; editing and saving the snippet retrains it immediately.

​
Set a sync frequency

Set the Auto Sync Frequency:
  • For a website crawl or sitemap import, in the Scrape Website or Add from Sitemap form (or later from Auto Sync Jobs).
  • For individual links, by opening the link in Links List, or selecting several and using Update Config.
Available frequencies depend on your plan. See plans for which schedules your plan includes.
Changing a frequency schedules future syncs; it does not start one immediately. To update content right now, select the links and use the bulk Resync button.

​
Sitemap scanning

Sitemap imports have a second schedule, Auto Scan Frequency: SiteGPT re-reads your sitemap.xml on that schedule and imports any new URLs it finds. This is how newly published pages join your chatbot automatically. Scanning discovers new pages; syncing refreshes existing ones. Use both for a fully hands-off setup.

​
Monitor your sync jobs

Open Auto Sync Jobs in the sidebar to see every scheduled job, its sources, and its run history. Open a job to see each run’s status and how many documents were updated.
Open the job’s run for details. The same causes that fail an import fail a sync: the site blocking crawlers, removed pages, or exhausted pages quota. Fix the cause; the next scheduled run picks it up, or resync manually.
Deleting an auto-sync job stops the schedule. Content that was already synced stays trained until you delete it from the list.