The document upload API is a two-step process. First, the metadata for the document is specified, and the API returns a file identifier so the file can be streamed directly to the next endpoint. You’ll need to look for the document_token returned in the response because that is necessary to pass to the next endpoint for streaming the file.