Attach a web source and crawl it now
POST /voice/collections/{collection_id}/scrape-sources
Crawl a site into this collection, now and then on a schedule.
THE FIRST CRAWL STARTS IMMEDIATELY. It does not wait for the cadence, so
creating a source with frequency: manual still crawls once. Up to
max_pages pages are fetched and embedded, and both cost money - which is
why this is billable and voice guards it with its EXPENSIVE rate limit.
frequency is closed to voice’s four cadences here. Voice itself takes a
bare string and only its scheduler notices an unknown one, so an invented
value there produces a source that quietly never re-crawls.
Authorizations
Parameters
Path Parameters
Collection id.
Collection id.
Header Parameters
Request Body required
A site to crawl into a collection on a schedule.
CREATING ONE CRAWLS IMMEDIATELY - it does not wait for the first scheduled
run. That is why this operation is billable rather than a plain write.
object
How often to re-crawl. manual never re-crawls.
Ceiling, not a target - a small site only crawls what exists.
Root URL. Everything under it is followed.
Responses
201
Successful Response
object
One voice object - a collection, an agent, a route, a chat.
Open-shaped: voice adds fields to these bodies on its own release cadence and a dropped-or-added key must not be a 500 in this service.
object
object
Whether there are more pages
Items per page
Current page number (1-indexed)
Total number of items
422
Validation Error