Per-client configuration

Content Intelligence

Auto-crawl runs the content extraction pipeline on pending pages after every successful connector sync. It does not run on a fixed clock — the trigger is the sync event itself, with a 60-second defer so the metadata writes commit first.

Auto-crawl

Opt this tenant in to platform-scheduled crawl + LLM extraction. Defaults to off.

Auto-crawl is disabled.

Fires 60 seconds after each successful connector sync. Capped per-run (default 100 rows) and protected by the daily LLM budget. Three consecutive failure-heavy runs auto-disable this flag — re-enable here once the upstream cause is fixed.

Last run

Same facts the connectors-page banner shows — kept here so the toggle and the evidence behind it live together.

Loading…

Add tenant website

Crawl the tenant's own public marketing site. Each subdomain is its own profile.

Tenant website profiles

Configured public-site crawl profiles. The cron picks up enabled profiles daily at 06:00 UTC.

Pick a tenant from the switcher to manage profiles.

Embeddings

Content embeddings power the learned content recommendations for people who aren't in a buying group. Tag extraction and embeddings use different providers, so tags can succeed while embeddings stay empty.

Loading…