This explains how to use the POST /pipelines endpoint to extract data from a PDF and return the data in JSON.

QuickstartWeb Extraction