Image OCR - Extract Text from Images in 40 Languages

Turn photos, scans, screenshots, receipts, invoices and documents into text. Paste a list of image links and get clean recognised text back, with a confidence score and, if you need it, the position of every line and word. No API key, no account at a third-party service, no per-minute quota.

What you get

• Recognised text for each image, with paragraphs and line breaks kept.
• Confidence score (0-100) per image, per line and per word.
• Coordinates: bounding boxes for lines and words (left, top, width, height in pixels) for redaction, highlighting and form parsing.
• 40 languages, including English, German, French, Spanish, Italian, Portuguese, Russian, Ukrainian, Polish, Turkish, Arabic, Chinese, Japanese, Korean, Hindi, Vietnamese, Thai and more. Mix languages with a plus sign, for example eng+deu.
• Auto-rotate for sideways and upside-down pages; EXIF orientation applied automatically.
• Image clean-up before recognition: grayscale, contrast, sharpening and enlarging of small images, which noticeably improves results on phone photos and screenshots.
• Bulk mode: hundreds of images per run, processed in parallel.

Typical uses

• Digitise receipts, invoices and business cards.
• Extract text from screenshots and memes for search and moderation.
• Read scanned contracts, forms and archives.
• Feed image text into LLM and RAG pipelines.
• Read labels, number plates and signs from photos.

Input

Field	Description
Image URLs	Direct links to JPG, PNG, WebP, TIFF, GIF or BMP files (required).
Languages	For example eng, eng+deu, rus+ukr.
Layout mode	Automatic, single block, single line, single word or sparse text.
Auto-rotate pages	Detect and fix sideways text.
Improve image before OCR	Grayscale, contrast, sharpen, enlarge.
Return lines / words with coordinates	Add bounding boxes.
Minimum word confidence	Drop uncertain words.

Example input:

{
    "startUrls": [{ "url": "https://tesseract.projectnaptha.com/img/eng_bw.png" }],
    "languages": "eng",
    "includeLines": true
}

Output

{
    "url": "https://tesseract.projectnaptha.com/img/eng_bw.png",
    "text": "Mild Splendour of the various-vested Night!\nMother of wildly-working visions! hail!",
    "confidence": 93.4,
    "language": "eng",
    "wordCount": 24,
    "lineCount": 3,
    "width": 600,
    "height": 130,
    "rotation": 0,
    "scrapedAt": "2026-10-14T08:00:00.000Z"
}

Pricing

You pay only for images in which text was found. Images without text and failed downloads are free.

Event	Price
Image with recognised text	$0.005 per image

Volume discounts apply automatically for Bronze, Silver and Gold Apify subscribers. 1,000 images cost about $5. Set a maximum cost per run to stay in control.

Tips for better accuracy

• Choose the right language. A wrong language is the most common reason for poor results.
• Keep Improve image before OCR on for phone photos and screenshots; switch it off for clean 300 DPI scans.
• For one line of text (a plate, a label) use Single line layout.
• For screenshots with many small labels use Sparse text.
• Handwriting is not supported; printed text works best.

Limits

• Links must return the image file itself, not a web page.
• Images up to 25 MB by default (configurable).
• Please make sure you have the right to process the images you submit.

FAQ

Is it free to try? New Apify accounts receive free monthly credit that covers hundreds of images.

Can I OCR PDFs? Use our PDF Text and Table Extractor, which includes OCR for scanned pages.

How is confidence calculated? It is the average of the recognition confidence of all words in the image.

Can I get the text in a file? Export the dataset as JSON, CSV, Excel, XML or HTML from the Storage tab or via the API.

More tools by the same author

• PDF Text and Table Extractor: PDF to text, tables and Markdown with OCR
• Bulk Image and File Downloader: Bulk download of images and files
• Video Frame Extractor: Extract frames and scenes from video
• Website Screenshot API: Full-page and mobile screenshots in bulk