Operator

Operator is now available as a third extraction base processor, extraction_operator. For long, table-heavy documents an extraction agent replaces the chunked pipeline: it plans, extracts repetitive structure programmatically, verifies totals and row counts against the source, and submits a schema-validated result. For other documents, the agent audits and corrects any fields the Review Agent flags with low confidence. Both paths require a single-file JSON Schema extraction.

API

Select it with "baseProcessor": "extraction_operator" on POST /extractors, POST /extractors/{id}, POST /extractors/{id}/versions, inline config on POST /extract and POST /extract_runs, and inline workflow extractor configs.
Settings Operator manages (Review Agent, multimodal, reasoning insights, array strategy, chunking) are applied automatically when a run is created.
Run usage.breakdown[].charges[] can now include a base charge with product: "extraction_agent" and unit: "usd" charges with product: "agent_usage".

Pricing

Operator runs bill a flat 2 credits/page with no extraction surcharges, plus Agent usage: the agent’s model provider cost converted to credits at your plan’s rate.

Read the Operator guide and Operator Pricing.

Strict Identifier Mode for Splitters

Splitters can now merge adjacent sub-documents that share the same extracted identifier into a single split, through the new advancedOptions.strictIdentifierModeEnabled option on the 2026-02-09 API version. Use it when one logical document repeats its identifier across its pages — a multi-page invoice whose invoice number appears on every page, for example — and you want one split per identifier rather than several consecutive splits of the same document.

Identifiers come from the identifierKey rule on the matching classification in splitClassifications when one is defined; without it the splitter falls back to a short descriptive identifier of its own, so define identifierKey on any classification you intend to merge by. The option is available on all splitting_performance versions and splitting_light >= 1.3.0

strict identifiers are on by default for splitting_performance >= 1.6.2 and splitting_light >= 1.4.1

See Splitter Configuration for details.