OCR: Scanned PDF to Text & Image to Text

Turn scanned PDFs and images into plain text, one row per page. Use it as a simple PDF to text or image to text converter for scans, faxes, photographed documents and screenshots. You pay per page, and pages that come back empty are free.

• Scanned PDFs and photos: PNG, JPG, TIFF (including multi-page), BMP, GIF and WEBP
• Digital PDFs: if a page already has a text layer, it is read directly instead of OCR'd, at a quarter of the price
• Free pages: blank pages, and pages where no text was found, are not charged
• Confidence score: each OCR'd page gets a score from 0 to 100, and low-scoring pages are flagged needsReview so you know which ones to check by hand
• Spreadsheet-ready: results are a flat table you can download as CSV, Excel or JSON from the run's Storage tab

How to use

1. Paste links to your files under File URLs. They must be publicly downloadable. Google Drive and Dropbox share links work if the file is shared with anyone who has the link.
2. Pick the languages printed in your documents. English is the default.
3. Click Start. When the run finishes, open the Output tab or export the dataset.

Pricing

What happens	Price
A page is OCR'd and text is found	$0.002 per page ($2 per 1,000)
A PDF page is read from its own text layer	$0.0005 per page ($0.50 per 1,000)
Starting a run	$0.00005 per run
A blank page, or a page where no text is found	free
A file that fails to download or isn't a PDF or image	free

Set Max total charge in the run options to cap your spend. The Actor stops cleanly when it reaches your limit.

Input

Field	Default	Notes
fileUrls	(required)	List of links to PDFs or images, up to 100 MB each
languages	["eng"]	English, Spanish, French, German, Portuguese, Italian, Dutch, Korean, Japanese, Chinese (Simplified)
ocrMode	auto	auto: OCR only pages without a text layer. always: OCR every page. never: text layer only
dpi	300	Rendering resolution for PDF pages before OCR (150 to 400)
maxPagesPerFile	200	Pages beyond this are skipped and not charged

Output

One dataset row per page:

{
  "fileUrl": "https://example.com/scan.pdf",
  "fileName": "scan.pdf",
  "page": 2,
  "pageCount": 3,
  "method": "ocr",
  "text": "This is a lot of 12 point text to test the\nocr code and see if it works on all types\nof file format.",
  "charCount": 120,
  "wordCount": 25,
  "ocrConfidence": 95.9,
  "needsReview": false,
  "error": null
}

method is ocr, text-layer or blank. A file that can't be processed gets a single row with error set, and the rest of your files still run.

Limits

• OCR is done with Tesseract. It works well on printed text and clean scans. Handwriting and low-quality phone photos give poor results, and those pages will show a low confidence score.
• Tables come out as lines of text, not as structured rows and columns.
• Password-protected PDFs are skipped.

Privacy

Files are downloaded into the run, processed, and not sent to any other service. Results stay in your own Apify storage, under your account's data retention settings.

About

This Actor was written and is maintained by an AI model (Claude, by Anthropic), with a human owner responsible for the account. If something breaks or a result looks wrong, open an issue on the Issues tab and it will be looked at.