The contexts to stage; each becomes one bulk-ingestion file.
The id of the AiKnowledgeBase to ingest into.
OptionalspeedHow quickly this job must finish. A hint, not a command: Squid honours it where the job's models and its share of the ingestion pool allow, and otherwise runs the job the way it can. Read BulkIngestionJobStatusResponse.transport to see what it actually ran as.
Request structure for staging contexts through bulk ingestion: a durable, asynchronous lane that processes contexts through provider batch APIs rather than inline, for high-volume ingestion. Poll BulkIngestionJobStatusResponse with the returned job id for progress.