Create Multi-URL Import
Create Multi-URL Import
Kick off an async multi-URL import. Accepts 1..N URLs in a
single job (capped per-deployment, default 50) and runs the
same per-URL pipeline as the sitemap importer. Returns 202 with
the import job row; client polls GET /{kb_id}/imports for
progress.
Authentication
Enter your API key with the Bearer prefix, e.g. ‘Bearer sk_…’.
Path parameters
Knowledge base id (prefixed external id, kb_...).
Headers
Request
Folder to import the documents into. Prefixed wire identifier
(kfolder_<26 char Crockford base32>); null/omitted = root.
Response headers
Response
The one lifecycle vocabulary shared by every async job (batch
calls, knowledge-base imports, agent-test runs, suite runs):
pending → running → a terminal state.
pending- accepted but not yet executing (queued for a worker, or deferred to a future scheduled time).running- actively executing.completed- ran to conclusion. The single terminal-success verb. For a job that produces a pass/fail judgment (an agent-test run), this means it produced a verdict - read the separateverdictfield for the judgment, not this status.failed- could not complete (an infrastructure or input failure), distinct from acompletedjob whoseverdictisfailed.cancelled- cancelled before reaching a natural terminal state.
Total URLs to process (the progress denominator).
JSON blob whose shape depends on kind — typically url,
max_pages, max_depth. Read it for display
only.