Extract text from an image with OCR
A sample scanned image read in English.
{
"urls": [
"https://raw.githubusercontent.com/lergassy/apify-actor-assets/main/document-text-extractor/sample-scan.png"
],
"language": "eng"
}Document OCR
OCR Text Extractor: turn images and scanned PDFs into text
Send image or PDF URLs, an uploaded file or base64 data and get the recognised text back, page by page, with an engine confidence score. Add word positions to build document parsers, or reuse a PDF's own text layer when it already has one.
★ 5.0 rating · 1 reviews18 users97.1% successful runs (30 days)
What data you get
text | Full recognised text of the file |
pages | Text of each page separately |
confidence | Engine confidence score per page |
words | Optional recognised words with confidence |
boundingBox | Optional position of each word on the page |
language | OCR language used |
fileName / url | Source of the processed file |
How it works
- Open the scraper on Apify and sign up (free).
- Enter routes, dates and options in the form.
- Click Start and download results as JSON, CSV or Excel, or call the API.
Pricing
$2.00 per 1,000 rows
Pay per event: a small charge per file processed plus a charge per page read with OCR. Use the page range and max pages settings to keep costs predictable.
Apify gives $5 free credits every month — enough to test this scraper at no cost.
Example tasks
Ready-made inputs. Copy one into the scraper form or send it through the API.
OCR a scanned PDF with word positions and confidence
First 3 pages, low-confidence words dropped.
{
"urls": [
"https://raw.githubusercontent.com/lergassy/apify-actor-assets/main/document-text-extractor/sample-scan.pdf"
],
"pageRange": "1-3",
"includeWordBoxes": true,
"minWordConfidence": 60
}Force OCR on a scanned PDF with automatic language detection
Ignore the PDF text layer, up to 20 pages.
{
"urls": [
"https://raw.githubusercontent.com/lergassy/apify-actor-assets/main/document-text-extractor/sample-scan.pdf"
],
"autoDetectLanguage": true,
"useExistingTextLayer": false,
"maxPagesPerFile": 20
}Batch convert images and PDFs to plain text
Two files, full text only.
{
"urls": [
"https://raw.githubusercontent.com/lergassy/apify-actor-assets/main/document-text-extractor/sample-scan.png",
"https://raw.githubusercontent.com/lergassy/apify-actor-assets/main/document-text-extractor/sample-scan.pdf"
],
"includePages": false
}Use cases
- Developers: add OCR to apps and document parsers without running your own engine.
- Back-office teams: pull text out of scanned invoices, receipts and contracts.
- AI and RAG builders: convert scanned PDFs into text for search and LLM pipelines.
- Researchers and archivists: digitise scanned papers and archives in many languages.
FAQ
Which languages are supported?
19 languages, including English, German, French, Spanish, Portuguese, Russian, Indonesian, Vietnamese, Arabic, Hindi, Chinese, Japanese and Korean. Script auto-detection is optional.
Does it work with scanned PDFs?
Yes. Scanned pages are read with OCR; pages that already contain a text layer can be copied directly, which is faster and more accurate.
Can I get word coordinates?
Yes. Turn on word positions to get every word with its confidence and its box on the page.
Is it legal to run OCR on documents?
Only process files you own or have the right to process. If documents contain personal data, handle the output in line with applicable privacy laws.
How do I integrate it?
Call it via the Apify API with URLs or base64 files, use the Python or JavaScript clients, or connect it to Make, n8n, Zapier or Google Sheets.