Image OCR (text from images, 20 languages)

Pull the text out of images in bulk: screenshots, scanned documents, receipts, photos of signs, product labels, memes and charts. Paste image links or upload an image. It supports 20 languages, including Chinese, Japanese, Korean, Arabic and Hindi (with vertical text for the first three), and gives a confidence score for every result. It runs on the open-source Tesseract engine, so you need no API key.

Price: US$0.0025 per image, plus a US$0.001 run fee. 100 images cost US$0.251.

What people use it for

• Data entry: text from scanned forms, invoices and receipts.
• Screenshot search: make screenshots and social media images searchable.
• E-commerce: read product labels, ingredients and packaging text.
• Content moderation and research: find text inside images at scale.
• AI pipelines: turn images into text an LLM can read, cheaply, before reaching for a vision model.

Example output

Illustrative values. Real output has the same fields.

fileName	text	meanConfidence	charCount	width	height
receipt-0412.jpg	CORNER CAFE\nLatte 4.50\nCroissant 3.20\nTOTAL 7.70	91.4	48	1080	1920
slide-07.png	Quarterly results\nRevenue up 12%	95.2	31	1280	720

{
  "sourceUrl": "https://example.com/receipt-0412.jpg",
  "fileName": "receipt-0412.jpg",
  "width": 1080,
  "height": 1920,
  "languages": ["eng"],
  "meanConfidence": 91.4,
  "charCount": 48,
  "text": "CORNER CAFE\nLatte 4.50\nCroissant 3.20\nTOTAL 7.70",
  "processedAt": "2026-10-27T09:00:00+00:00"
}

With includeLines on, each item also has lines: every line of text with its own confidence score. Paragraphs are separated by a blank line in text.

Input

Field	What it does	Default
imageUrls	Direct links to images, up to 1,000. JPEG, PNG, WebP, GIF, BMP, TIFF. 30 MB and 60 megapixels per image.	1 sample image
uploadedFile	Or upload one image from your computer.	none
languages	Up to 4 of: English, German, French, Spanish, Italian, Portuguese, Dutch, Polish, Turkish, Russian, Ukrainian, Arabic, Hindi, Indonesian, Vietnamese, Thai, Japanese, Korean, Chinese (Simplified), Chinese (Traditional), plus vertical-text Japanese, Korean and Chinese.	English
layout	auto for documents and screenshots, single-block, single-line for labels, sparse for signs and photos.	auto
includeLines	Add line-by-line output with confidence scores.	false

Small images are enlarged before OCR, and phone photos are rotated using their camera data, both of which improve accuracy.

Pricing

Event	Price
Run fee, once per run (per GB of memory; the default is 1 GB)	US$0.001
Image read	US$0.0025

Examples: 1 image costs US$0.0035. 100 images cost US$0.251. 1,000 images cost US$2.501. Images that fail to download or open are listed with the reason and not charged.

Limits

• Printed text works best. Handwriting is read poorly.
• Old newspapers and dense multi-column pages, especially in Chinese or Japanese, are read poorly. Modern printed text and screenshots work well.
• Very stylised fonts, text on busy photos and tiny text lower accuracy. meanConfidence below about 60 means the result needs checking.
• Pages behind a login cannot be fetched.

FAQ

How accurate is it? It is best on clean printed text such as screenshots and scans. Photos, small text and low-resolution images do worse. Check meanConfidence to spot results that need a second look.

Should I pick more languages to be safe? No. Pick only the languages in your images. Each extra language makes OCR slower and less accurate.

Can it read PDFs? Use PDF to Text and Markdown below. It reads the text layer and uses OCR only on scanned pages.

Can an AI agent call it? Yes. Pass image links, get compact JSON back, and pay per image.

More tools from the same developer

Same style: clean JSON output, pay only for results, no charge on failed items.

Tool	What it does
Bulk Keyword Search Volume	Google Ads search volume, CPC, competition and 12-month trend for up to 10,000 keywords
Google Trends API	Interest over time, by region, and rising queries and topics for up to 500 keywords
Google SERP Results	Organic results, featured snippets and People Also Ask for any keyword, country and device
Google AI Overview Tracker	Which keywords show an AI Overview, what it says, and whether it cites your site
Audio and Video Transcriber	Timestamped transcripts from audio and video files in 90+ languages
Subtitle Generator	Ready-to-use SRT and WebVTT subtitles from audio and video files
Podcast Transcriber	Timestamped transcripts of podcast episodes from any RSS feed or Apple Podcasts link
PDF to Text and Markdown	Clean text and Markdown from PDFs, with OCR for scanned pages
Bulk PageSpeed Insights Audit	Core Web Vitals and Lighthouse scores for many pages, mobile and desktop