PDF & Image to Text (OCR in 21 languages)

Turn PDFs and images into clean text or Markdown, ready for search, spreadsheets, LLMs or RAG pipelines.

• OCR only where it's needed. Pages with a real text layer are read directly and charged at the cheap text-page price. Only scanned pages (pages that are images with no usable text) go through OCR.
• Non-Latin OCR included. Arabic, Urdu, Persian, Sinhala, Tamil, Hindi, Bengali, Malay, Indonesian, Turkish, Chinese, Japanese and Korean, plus major European languages, using the open-source Tesseract engine. Mix languages in one document. Accuracy varies by script and scan quality, so test a few pages first.
• PDFs and images in one run. PDF, PNG, JPEG, TIFF (including multi-page), GIF, BMP and WebP.
• Clear results for every file. Each file returns its text, page counts, how each page was read, and a plain-English error if it couldn't be processed. One bad link never stops the rest of the run.

How to use

1. Paste direct links to your files under Files. Each link must download the file itself.
2. Pick the OCR languages used in your documents (default: English).
3. Run. Download the results as JSON, CSV, Excel or via the API.

Input

Field	What it does	Default
Files	Direct links to PDFs or images	–
OCR mode	auto OCRs only scanned pages; always OCRs every page; never uses the text layer only	auto
OCR languages	Tesseract codes: eng, ara, urd, fas, sin, tam, hin, ben, msa, ind, tur, fra, deu, spa, por, ita, nld, rus, chi_sim, jpn, kor	eng
Output format	text or markdown (a title plus one heading per page)	text
Include per-page text	Adds each page's text and how it was read	on
Max pages per file	Pages beyond this are skipped and not charged	200
Max file size (MB)	Larger files are skipped with an error	50

Output

One item per file:

{
  "url": "https://example.com/report.pdf",
  "status": "ok",
  "fileType": "pdf",
  "pageCount": 12,
  "pagesProcessed": 12,
  "truncated": false,
  "textPages": 10,
  "ocrPages": 2,
  "emptyPages": 0,
  "ocrLanguages": ["eng", "ara"],
  "format": "text",
  "characters": 18234,
  "text": "…",
  "pages": [{ "page": 1, "method": "text", "text": "…" }]
}

Files that can't be processed come back with "status": "error" and a reason in error (not found, too large, password-protected, not a PDF or image). You aren't charged for failed files.

Pricing

You pay per page actually processed, plus a tiny fee per run:

Event	Price
Text-layer page	$0.001
OCR page	$0.004
Run start (once per run)	$0.00005

Example: one run with a 20-page PDF that has 2 scanned pages costs 18 × $0.001 + 2 × $0.004 + $0.00005 = $0.02605. Set a maximum cost per run in Apify and the Actor stops cleanly when it's reached.

Typical uses

• Make scanned contracts, invoices and reports searchable.
• Prepare PDFs for LLM or RAG pipelines (Markdown output, one heading per page).
• Digitise Arabic, Urdu, Sinhala or Tamil documents, which English-only OCR tools can't read at all.
• Bulk-convert a list of document links to JSON, CSV or Excel.

Good to know

• Links must be publicly downloadable. Pages that need a login, and links to private or local network addresses, are refused.
• OCR quality depends on the scan. Clear scans at 200–300 DPI work best.
• Only process documents you have the right to use.

Acceptable use

Please don't use this Actor for, or on behalf of, gambling or betting, alcohol, adult or immodest content, interest-based lending, or deceptive marketing. It's built for ordinary business, research and personal use.