This endpoint uploads a given document and creates a documenttranscription task. In this task, our workers will look at single or multiple page documents in order to transcribe and annotate any information you would like. Example use cases include transcribing invoices from PDFs, manually scraping a screenshot for information, etc.
In the response format, annotations can have references to each other to handle use cases that require structured data, such as key and value linking, table header and table data linking, or for handling content that spans multiple pages. We currently support application/pdf, image/png, image/tiff, image/jpeg, image/jp2, image/gif, image/webp, and image/bmp. We accept the file contents as the request body. Do not set the is-base64 header unless you're base64 encoding the contents of the file.