Source documents can be archived at ingest
Docparser can retain the exact bytes used by a document job so an eligible result can later become a golden template.
Docparser now archives source bytes when a document job ingests them. This avoids re-fetching a URL that may have expired or changed after the original extraction.
Archiving runs independently of extraction: a failed archive upload logs a warning without failing the document job. Single-document jobs record the archived object key, while multi-part inputs keep each part under the job’s archive prefix.