Submit documents
Queue a document (or several about one patient) and a prompt for summarization. Returns immediately; poll GET /v1/jobs/{id}, then fetch the result.
Authentication
Customer API key or OAuth 2.0 access token (scope summarization:process)
Headers
Optional, recommended: retrying with the same key and body returns the original job instead of a new one (per organization, 24 h). 1-128 of [A-Za-z0-9_.:-]
Request
One document. Send it alone, or instead use any of documents, prep_documents and nlp_results (they can be combined)
Several documents about one patient, summarized as one: a single summary that cites them. At most 50, each at most 64 KiB, 1 MiB in total
Documents as Prep produces them: the objects of its JSONL file (one per line), sent as an array. Each is summarized like an item of documents, with its id, category, chart date and subject details. Same limits as documents
NLP API results you already have (emtellipro-json-2, as returned, including the documents’ text); their documents are summarized. They have no dates or patient details; add those with document_metadata
FHIR R4 Bundles (e.g. from the Synth API). Each one’s clinical notes (DocumentReference with a text attachment) are summarized, along with its structured record — problems, medications, procedures, immunizations, allergies, encounters, laboratory results and vital signs — and the patient’s age and sex. Notes that can’t be read (e.g. PDF attachments) are skipped and named in a documents_skipped warning
Chart dates and subjects by document id (or position, for documents without one), over any of documents, prep_documents, nlp_results and fhir_bundles
Response
Idempotent retry: the job already created with this Idempotency-Key
expired: the result was downloaded or the 24-hour window passed
UTC, YYYY-MM-DDTHH:MM:SSZ
Present when status is failed
While running, for jobs with several documents

