Batch
Asynchronously execute batch jobs of up to 100 tasks across a concurrent worker pool with partial-failure isolation and job polling.
Reduce round-trips for indexing pipelines. Batch jobs run asynchronously: submit up to 100 tasks in a single job with POST /v1/batch and poll status with GET /v1/batch/{id}. Tasks are executed concurrently across a worker pool (default 10, maximum 20 workers). Credits are charged at 1 credit per successful task (both for transcript and basic_info; while individual single basic-info requests are free, batch processing consumes high-concurrency cluster capacity). Failed tasks consume 0 credits.