Japanese OCR for screenshots and photos
Upload or paste an image and copy the Japanese text as plain lines.
Your screenshot
or drop an image here, or press Ctrl+V to paste one. PNG, JPEG or WebP. Large images are shrunk in your browser before they are sent.
Text
Short answer
Japanese OCR turns the writing in a picture into text you can copy. Upload or paste an image above and this page returns the Japanese lines in reading order, with no extra analysis. Each OCR run uses one of your three free daily scans, shared with the main screenshot tool.
Japanese OCR options compared
The right choice depends on the source image and on what you do with the text afterwards.
| Option | Output | Best for | Watch out for |
|---|---|---|---|
| This page | Plain Japanese lines | Screenshots, game text, printed pages | 3 free scans per day, shared quota |
| The main scan tool | Text plus furigana, word table, grammar, translation | Study and reading practice | Same daily quota |
| Generic OCR websites | Plain text | Scanned documents in many languages | Japanese accuracy varies, ads and limits differ |
| Google Lens | Selectable text and translation | Quick look-ups on a phone | No reading aids for learners |
| Phone camera text selection | Selectable text in the photo app | Signs and menus | Depends on the phone and OS version |
How to run Japanese OCR on an image
- Add the image. Drop a file on the box, choose one with the button, or press Ctrl+V to paste a screenshot from the clipboard.
- Run OCR. Press the scan button. The browser shrinks the image first, which keeps uploads small.
- Copy the text. Press Copy text to put every line on the clipboard. You can also select part of the result by hand.
- Proofread. Compare the lines with the picture, especially similar kanji and small kana, and fix any slips.
What this OCR does and does not do
This page returns text only. It does not add readings, meanings or a translation. That keeps the request small and fast. If you want those, use the screenshot scan on the home page, which costs the same single scan.
The text is read by an AI vision model, not by a classic OCR engine. That helps with game fonts and mixed backgrounds, but the model can also invent or skip characters. Always proofread.
Tips for better Japanese OCR
- Crop to the text. Less artwork around it means fewer mistakes.
- Use a straight, sharp image. Blur and strong skew hurt small kana.
- Vertical text is supported less reliably than horizontal text.
- Look closely at pairs such as ソ and ン, シ and ツ, and 未 and 末. These are common slips.
- Furigana printed beside kanji may be read as extra text. Delete it if it appears.
When you already have the text
Once you have the lines, paste them into the furigana converter to add readings for free. The converters run in your browser and do not use a scan.
Frequently asked questions
Is this Japanese OCR free?
Yes, up to 3 scans per day per visitor. OCR runs, screenshot scans and sentence explanations share that quota. Paid packs are planned but not on sale yet.
Does it work on handwriting?
Neat handwriting sometimes works, but results are unreliable. Print, subtitles and game text are much better. Check the output against the image either way. Block capitals written slowly do better than cursive or fast notes.
Is the image kept?
No. It is processed in memory and not stored here. It is sent to an AI model provider through OpenRouter to read the text. The privacy page lists what is logged.
What file types can I upload?
PNG, JPEG and WebP. Your browser scales the image to at most 1600 px on its longest side and the upload must be 4 MB or smaller after that.
Why did the OCR miss some text?
Small, stylized or low contrast text is the usual cause. Crop closer, use a larger screenshot, or try again with a cleaner capture. A failed scan that returns an error does not use up your daily quota.
Last updated: