Bulk CV intake

Bulk intake is the entry point for high-volume hiring. It accepts a set of documents, extracts structured candidate records, and hands them to the ranking pipeline. This page describes what it accepts, how fast it runs, and where it will stop.

Supported formats

PDF, DOCX, DOC, RTF, TXT and ODT are parsed natively. Images and scanned PDFs are routed through OCR. Archives are unpacked one level deep; nested archives are ignored rather than rejected, and appear in the skipped list with a reason.

Throughput and limits

Standard throughput is approximately 900 documents per minute. Enterprise workspaces get burst capacity to 2,400. A single batch may contain up to 20,000 documents; larger sets are accepted and split automatically into sequential batches, processed in windows, with an email on completion.

Pre-flight validation

When enabled, the batch is inspected before processing begins. Unreadable, encrypted, zero-byte and oversized files are reported up front, and you decide whether to proceed, exclude them, or fix the source. Pre-flight adds roughly 4% to total run time and removes most mid-run failures.

Deduplication

Candidates are matched on email first, then on a fuzzy key of name plus phone plus most recent employer. An exact email match merges silently. A fuzzy match is flagged for review rather than merged, because a false merge is far more expensive to undo than a duplicate is to keep.

Retention

Source documents are retained for the period set in workspace policy, 24 months by default, and are then deleted irreversibly along with derived artefacts. Candidates exercising a deletion right are removed within 30 days across all batches, including from completed audit bundles where the law requires it.