PDF Extractor — PDF to text, tables & metadata

For developers, analysts and AI/RAG pipelines that need the contents of PDFs as clean data. Give it public PDF URLs and get per-page text, tables as rows and cells and document metadata (title, author, producer, page count). A pdf to text and pdf table extractor in one Actor, with no OCR and no files to upload. $0.004 per PDF (first 20 extracted pages included) + $0.0005 per extra page. Leave the input empty for a free demo.

Use cases

• Feed reports, filings, manuals and papers into a RAG pipeline or LLM, page by page with page numbers
• Pull tables out of financial reports, price lists and government statistics into JSON/CSV
• Build a searchable text index of a PDF library, with title and author from the metadata
• Spot scanned documents (image-only pages) before sending them to an OCR service

How it works

• Each PDF is downloaded once, then parsed with pdfplumber (MIT licence). Text and tables come from the PDF's own text layer.
• pages limits the range (1-5,8, 3-), maxPagesPerPdf caps long documents (default 200, max 2000).
• Two output shapes: perPage (one row per page) or perDocument (one row per PDF with all text and tables).
• Scanned pages: a page with no text layer and only images is returned with scanned: true and empty text. There is no OCR.
• Pay only for results: a page is charged only when text or tables were extracted. Failed downloads, unreadable PDFs, blank pages and scanned pages are free. The free demo (empty input) is not charged.
• Safe by design: URLs that resolve to internal addresses (localhost, private ranges, link-local such as 169.254.169.254) are never contacted, including redirect targets (blocked_private_address). Max 5 redirects, 50 MB per file, 60 s per download, %PDF signature required.
• Per-document SUMMARY in the key-value store: title, author, producer, pageCount, pagesExtracted, pagesCharged, scannedPages, error.

Input

{ "urls": ["https://example.com/report.pdf"], "pages": "1-5,8", "extractTables": true, "outputMode": "perPage", "maxPagesPerPdf": 200 }

Output (perPage)

{
  "url": "https://example.com/report.pdf", "page": 1, "text": "Quarterly Report ...", "charCount": 1834, "scanned": false, "error": null,
  "tables": [{"header": ["Region", "Revenue", "Growth"], "rows": [["North", "1,200", "5%"], ["South", "900", "7%"]]}]
}

header is the first row of a table when it looks like a header (all cells filled, no numbers), otherwise null and the first row stays in rows. error is e.g. http_404, not_a_pdf, file_too_large, timeout, too_many_redirects, encrypted, unreadable_pdf, blocked_private_address, bad_pages:... (a failed PDF produces one row with page: 0). In perDocument mode each row has text, tables (each with its page), title, author, producer, pageCount, pagesExtracted, scannedPages, error.

Pricing

$0.004 per PDF (pay per event, pdf; includes the first 20 extracted pages) + $0.0005 per extra page (page). Example: a 10-page PDF costs $0.004, a 100-page PDF $0.044. Failed, scanned-only and empty PDFs are free. For comparison, other PDF text Actors in the Apify Store charge about $0.003 to $0.02 per PDF (per document) or $0.003 to $0.005 per page.

Limits and notes

• Text order and table detection follow the PDF's layout; complex multi-column pages and borderless tables may need checking.
• Password-protected PDFs are not opened (encrypted). Only the PDFs you list are downloaded; links inside them are not followed.
• Only process PDFs you are allowed to download. The text belongs to the document's rights holder.

More tools from the same developer

• Website Screenshot API — screenshots and PDFs of web pages: full page, mobile, element, no cookie banners
• Google Sheets Export — export an Apify dataset to Google Sheets