Japanese image and PDF to text
Japanese (日本語) is written in the Japanese script. u2ocr reads it with a model trained on printed Japanese text, which downloads once into your browser and is kept for your next visit.
- Japanese ready to go
- Images and PDFs
- Powered by i2OCR
日本語
Japanese
Drop images or PDFs here
Photos, screenshots, scans and multi-page PDFs. Add one file or a hundred.
PDF, PNG, JPG, WEBP, GIF, BMP · up to 200 MB per file · or paste a screenshot
Japanese tips
Getting accurate Japanese text
What to keep in mind when you scan or photograph Japanese documents.
Vertical columns
For text set in vertical columns, choose Japanese (vertical) instead. It follows the top-to-bottom reading order.
Choose Best for small characters
Fine strokes merge together at low resolution. Use the Best setting for footnotes, captions and small print.
Paragraphs without extra spaces
When you join broken lines into paragraphs, u2ocr does not add spaces between characters, so sentences stay intact.
Mixed with English?
Documents that switch between Japanese and English, such as forms or technical papers, read better with both languages selected.
FAQ
Questions, answered
About languages, accuracy, privacy and cost.
Which languages are supported?
118 languages, including English, Spanish, French, German, Hindi, Arabic, Chinese, Japanese and Korean, plus vertical Korean, and historical scripts like German Fraktur. Choose one recognition language per run.
How do I get the best results?
Start with a sharp, well-lit and straight image, choose every language that appears in it, and pick the Best resolution for small print. For receipts, signs and forms, choose "Scattered text" as the page layout.
Are my files uploaded anywhere?
Yes. Images and rendered scan pages are sent to i2OCR for recognition. Digital PDF text can be extracted in your browser.
Is u2ocr really free?
Yes. There is no account or subscription. i2OCR service limits apply.
Have Japanese text to copy?
Drop your images or PDFs into the tool above. Japanese is already selected.
Extract text now