Standardize a single document using the V3 agentic engine. V3 uses an agentic architecture that automatically determines the optimal extraction strategy for each document - no manual configuration of display modes, split modes, or effort levels needed.
Unlike V2, this endpoint processes a single document per call. To standardize multiple documents, fire parallel requests.
If no schemaId is provided, a draft schema is first generated from the document, using guidelines (if sent) as instructions, then the document is standardized against it. The generated schema's ID is returned in the response's schemaId field. Generating the draft costs 1 credit per page on at most the first 5 pages, in addition to the standardization credits.
Parameters
schemaId- Optional schema to standardize against. If omitted, a draft schema is auto-generated from the document.pages- Optional list of 0-indexed page numbers to standardize. If not provided, the entire document is standardized.guidelines- Optional extraction guidelines. They replace the schema's saved guidelines for this call. Without aschemaId, they also steer the draft schema.useMetadata- Whether to include document metadata during extraction (default: false).effortLevel-standard(default) uses faster models,highuses the best models for better results at +2 credits per page.
Note: The guidelines field has a maximum length of 50,000 characters.