Image & Scanned PDF OCR - Extract Text from Images

Extract the text of images and scanned PDFs in bulk with OCR: screenshots, photos of documents, scanned invoices and contracts, receipts, old PDFs without a text layer. One row per image or PDF page, with line breaks kept, word count and a confidence score.

• Images and PDFs. PNG, JPG, WEBP, TIFF (all pages), GIF, BMP and PDF files (each page rendered at 300 DPI), up to 30 MB per file.
• 8 languages. English, French, German, Spanish, Italian, Portuguese, Dutch and Polish, alone or combined.
• Readable output. Lines and paragraphs are kept as in the document, so tables of figures and addresses stay readable.
• Cheap and predictable. $2 per 1,000 pages, where most OCR Actors charge $3 to $15. Pages without text are not charged.
• Safe. Files are processed in memory and never shared. Only public URLs are accepted: internal network addresses are refused.

What can I use it for?

• Invoices and receipts: turn scanned invoices into text, then extract amounts and dates with your own rules or an LLM.
• Scanned PDFs for RAG: make old scanned reports searchable and feed them to a vector database.
• Screenshots and images: read the text of product images, banners, slides or charts.
• Archives: digitize scanned letters, forms and records in bulk.
• AI agents: let an agent read any image through the Apify MCP server (see below).

PDFs that already contain text (exported from Word, for example) are read faster and more accurately by PDF, Word, Excel & PowerPoint to Markdown. Use this Actor for scanned pages and images.

Input

{
    "urls": [
        "https://raw.githubusercontent.com/naptha/tesseract.js/master/tests/assets/images/testocr.png",
        "https://example.com/scans/invoice-2026-001.pdf"
    ],
    "languages": ["eng", "fra"],
    "maxPagesPerPdf": 30
}

Output

One row per image, or per page for PDFs and multi-page TIFFs.

Field	Meaning
url	The file
fileType	image or pdf
page / pageCount	Page number and total pages (null for a single image)
text	The text, with line breaks; paragraphs separated by a blank line
wordCount	Words read
confidence	Mean OCR confidence of the words, 0-100. Below 60, check the image quality
width / height	Size of the page image, in pixels
error	Why a file could not be read (not charged)

Pricing

$0.002 per page read ($2 per 1,000 pages) plus Apify's small run-start fee. No subscription, lower prices on higher Apify plans. Pages without text, invalid URLs, download errors and unsupported files are free.

Tips for the best results

• Use the right language: accents and special letters are read much better.
• Scans at 200-300 DPI give the best results; tiny or blurry photos give low confidence.
• Handwriting is not supported well: this is printed-text OCR (Tesseract).

Use it from AI agents (MCP)

Add image-ocr to the Apify MCP server in your MCP client (Claude, Cursor, VS Code...).

Integrations

Apify API, schedules, webhooks, Make, Zapier, n8n. Export as JSON, CSV, Excel or HTML.

FAQ

Which OCR engine is used? Tesseract 5, the open-source engine originally developed by HP and Google.

Something wrong or missing? Open an issue on the Issues tab.