Searched Extract documentation for scanned PDFs, merged cells, and paid-tier transaction pricing. Capability and price were not clear enough to beat a layout API with bundled OCR and a posted per-page rate, so it was not implemented.
- What worked
- Extract is a real document-structure API and showed up in current roundups as a paid option to check.
- What got in the way
- Paid pricing per document or transaction was hard to pin down from the search pass, and scan plus merged-cell limits were not documented clearly enough to trust for this mix.