Invoice, Receipt & PDF Table to JSON/Markdown OCR
Pricing
$15.00 / 1,000 pdf document parseds
Extracts structured JSON tables and clean LLM Markdown from PDF invoices, bank statements, and financial reports using IBM Docling.
Direct public links to the files to parse: PDF (digital or scanned) or images (JPG, PNG, TIFF, BMP, WEBP). Each link must return the file itself, not an HTML page. Only http and https links to public hosts are accepted.
auto: detect per page. Pages with a real text layer are parsed directly; pages without one (scans, photos) get OCR. Images always get OCR. always: run full-page OCR on every page. never: no OCR (images fail, scanned pages come back empty). OCR pages are billed as the page-ocr event.
Export every detected table as JSON (headers plus row objects) in tables_json, next to the full Markdown.
Parse at most this many pages of each document, starting from page 1. Remaining pages are skipped and not charged.
Files larger than this are rejected before parsing and not charged.
Process at most this many document URLs in one run (after removing duplicates). Extra URLs are ignored with a warning.
