Generation lookup
Fetch the record of one request by its generation id.
Every response carries a generation id in x-batchwell-generation-id and in usage/batchwell fields.
curl
The record contains the requested and answering model, provider, status (completed, failed, blocked,
cancelled or pending), token counts, charge_micro, time to first token, latency and each upstream attempt.
It never contains prompt or completion text: we do not store it.
You can only look up generations that belong to the key's organization. The dashboard's Usage page shows the same records.
usage_source is upstream when the provider reported token counts, or estimated when it did not and we counted
delivered tokens ourselves.