Upload a document to the inbox

Adds one PDF, JPEG or PNG to the document inbox in a single call, its bytes sent as standard base64 in content_base64 (at most 700 KiB decoded; production requests are capped at 1 MB on the wire). The document is read, matched to a vendor and checked for duplicates exactly like a file uploaded in the app, and is marked with source api. Answers 202 with status: accepted and the processing_id to poll with documentExtraction.processing.status, or 200 with status: duplicate and duplicate_of naming the document already in the inbox (nothing new is created), or 200 with status: failed and a reason_code (retry with a new Idempotency-Key). Base64 may be line-wrapped. An optional sha256 is checked against the decoded bytes. Use an upload session for larger files. Idempotency-Key is required. Requires the document-extractions:write OAuth scope. The credential's user needs the create_document_extractions permission.

Recent Requests
Log in to see full request history
TimeStatusUser Agent
Retrieving recent requests…
LoadingLoading…
Body Params

Fields accepted to upload a document to the inbox. The schema identifies required and optional values.

string
required
length ≤ 255

As the inbox shows it, e.g. invoice-1042.pdf.

string
enum
required
Allowed:
string
required
length ≤ 955736

Standard base64, no data: prefix; at most 700 KiB decoded, line breaks not counted. Larger files: an upload session.

string | null
^[A-Fa-f0-9]{64}$

Hex SHA-256 of the file; refused when it does not match.

Headers
string
required
length ≤ 191

Required unique key whose request hash and terminal response are stored durably for 24 hours.

Responses

Language
Credentials
OAuth2
Missing 1 required scope
LoadingLoading…
Response
Click Try It! to start a request and see the response here! Or choose an example:
application/json
*/*