PDF to Markdown & Tables — CSV, Invoice Fields · $0.03/page

PDF to Markdown with tables as CSV: extract text, headings and tables from PDF and DOCX files, plus invoice or receipt fields (number, dates, VAT, total) by template or your own labels. Scans are read with OCR. Rule-based, no AI model. Failed files are free, up to 20 files a run.

Turn PDF and DOCX files into clean Markdown with headings, lists and tables, every table also as CSV rows, and optionally named fields (invoice number, total, dates) found by label. Layout rules, no AI model: the same file gives the same result. Per page pricing, files fetched from a URL, up to 20 files per run. Scanned pages are read with OCR (Tesseract) and billed separately.

What you get

Structured JSON for each run, ready for spreadsheets, alerts and AI agents.

Input

Field	Type	Default	Description
urls	array		Public URLs of PDF or DOCX files (up to 20).
maxPagesPerFile	integer	50	Stop after this many pages per PDF. Each processed page is one billable result.
fieldHints	array		Optional fields to find, one per line as name
template	string (invoice, receipt)		Ready-made field set: invoice (number, dates, supplier, customer, VAT number, subtotal, VAT, total, currency) or receipt (merchant, date, total, VAT, payment method). Combine with fieldHints to add your own fields.
ocr	string (auto, off)	auto	auto reads pages without a text layer (scans) with OCR, billed as ocr-page; off skips them.
ocrLanguages	string	eng	Tesseract language codes joined by +, for example eng+dan+deu.
includeTablesCsv	boolean	true	Add every detected table as CSV text next to the rows.

Example input

{
  "urls": [
    "https://www.rfc-editor.org/rfc/rfc9110.pdf"
  ],
  "maxPagesPerFile": 3
}

Example output (sample run recorded 2026-10)

{
  "files": 1,
  "extracted": 1,
  "pagesProcessed": 4,
  "billableResults": 4,
  "datasetItems": [
    {
      "url": "https://www.rfc-editor.org/rfc/rfc9110.pdf",
      "ok": true,
      "kind": "pdf",
      "pages": 4,
      "characters": 5134,
  …

Pricing

Event	Price
page (each successful result)	$0.03
OCR page (scanned page read with OCR)	$0.06

• A DOCX file is counted as one page per 3,000 characters of extracted text.
• Up to 20 files per run and up to 200 pages per PDF (default 50); pages read with OCR are billed as OCR pages.
• Template fields cover English, German, French, Spanish, Dutch and the Nordic languages; fields that are not found are reported as not found.

Failed runs are not charged. Free trial credits from Apify cover your first runs.

FAQ

Is it AI? No. It is rule-based layout analysis, so the same file gives the same result. Scanned pages are read with OCR (Tesseract) and flagged in the output.

Can it find invoice fields without setup? Yes. Set template to invoice or receipt: it looks for number, dates, supplier, customer, VAT number, subtotal, VAT, total and currency in English, German, French, Spanish, Dutch and the Nordic languages. Each field shows the line it came from and a confidence level; a field that is not found is reported as not found, never guessed.

What about complex layouts? Merged cells, multi-line headers and scans can need checking; the output flags OCR pages. Tables are returned as rows and as CSV.

Are my files stored? Files are fetched for the run and not kept beyond Apify's own run storage.

Where do I report a problem? Open an issue on this Actor's page; replies are written by the Ambolt service team, usually within a day.

Review

If this Actor saved you time, a short review on its Store page helps others find it and tells us what to improve.

Keywords

PDF to Markdown, PDF table extraction, PDF to CSV, PDF to JSON, DOCX to Markdown, invoice data extraction, document parsing, RAG, data entry automation, document to JSON

Notes

Data comes from the source named in each result and keeps its licence and attribution. Informational only, not financial, legal or tax advice. Open an issue on the Actor page for questions; replies come under the product name.