bunkoOCR
OCR with DeepLearning
Free · In‑App Purchases
This is an Optical Character Recognition(OCR) software with deeplearing model. All computation run on the device, no data is sent outside. This app can recognize horizontal and vertical text with ruby in facing page on Bunko(pocket edition) book. This app has best performance with Japanese and English text.
more - Add import images from pdf file.
- Add export OCR results as searchable pdf file.
- Fix transformer process to reduce OCR error.
- Add edge crop of image to process OCR.
- Add rotate 90 degrees direction and auto rotate detection.
3.15 08/23/2025
- fix crash bug in OCR processing some files.
3.14 06/27/2025
- fix Transformer model for context correction.
- fix bug not display some characters in OCR result.
- fix bug on indent detection.
3.13 06/24/2025
- Fix bug in OCR result view.
- Fix edit window of JSON result.
- Fix bug of Text-to-Speech in OCR result.
3.12 05/20/2025
- Fix bug in OCR result window.
3.11 05/17/2025
- Fix bug in search function on OCR result view.
3.10 05/13/2025
- Fix search function on OCR result view.
- Add function to edit JSON file of OCR result data and export as training data for machine learning.
3.9 05/12/2025
- Fix crash bug in some condition.
- Improve UI.
3.8 04/30/2025
- On older devices, fall back to CPU mode when crash loading ML model.
- Fixed incorrect recognition of "warichu"; annotation in main text by two line and "tate chu yoko"; horizontal text in vertical text.
- Fixed some processing errors.
3.7 04/29/2025
- Add option of combine distance for one line.
- Fix option of line search direction; auto, horizontal only, vertical only, horizontal main, vertical main.
- Add option default font size and ignore new lines for OCR result window.
- Show indent or line end combine marker in OCR result overlay window.
- Ignore small ink spot on paper as misidentified ruby.
- Fix logic of line indent.
- Add text search in OCR result window.
3.6 04/28/2025
- Fix crash bug when some image in process.
3.5 04/15/2025
- Add option wait sleep to prevent high-cpu kill.
- Fix unnecessary reload in result view.
- Fix detection logic for text indent.
- Add option to select pages for re-analyze.
- Add option to export images for other applications.
- Fix button caption for VoiceOver.
3.4 04/14/2025
- Fix jump bug the two digits section title.
- Fix bug in json export.
- Fix bug in remove new line mode.
- Add option to add extra newline on the page and segment boundary.
- Add support to open archive file.
- Add option to sort by filename.
- Add option threshold for block indent.
- Fix page edit menu.
- Add option that putting dummy page when remove the page.
3.3 04/08/2025
- Improved camera capture.
- Improved the process of searching for sentence connections. Fixed bug of sentence ordering.
- Added an option to ignore parts other than the main text. This is a test implementation and is not complete.
- Added an option to remove line breaks for display and export.
- Added an option combine all pages and display continuously.
- Fixed a bug in the text-to-speech.
- Added an option scroll to follow reading position in text-to-speech.
- Added an option to correct OCR mistakes through replacing text.
- Added a keyboard shortcut for page scrolling.
3.2 04/02/2025
- fix export menu bug
3.1 03/12/2025
- update ML model
3.0 03/11/2025
- Update ML model for more accuracy.
- Add option to ignore small text blocks(page number or subtitle).
- Fix text block order.
2.0 01/05/2024
- Improve camera capture
- Add crop option on camera capture.
- Add 48MP capture option if available.
- Focus score information to focus easier.
- Fix bug in OCR process
1.14 10/03/2023
- fix crush bug on iOS 17.
1.13 09/21/2023
- Add image pre-process options.
1.12 09/18/2023
- fix UI bug.
- add purchase menu. all app functions are still free.
1.11 09/13/2023
- Add function to process images from clipboard.
- Add function to capture with camera repeatedly.
- Add option to select delay timer time of camera.
- Add function to edit and add pages.
1.10 09/04/2023
- Add bookmark. Restore the page and position previously read.
- Add configuration of text result color.
- Add export text as html ruby tags.
1.9 08/27/2023
- Fix bug in ruby recognition process.
- Horizontal ellipsis(…) in Japanese text is not correctly drawing in some sentence, add a option fix this issue.
"Force Japanese font shape" is default on. If you want process English text, this option should set off.
- Set default value of "detection cutoff" to 0.5. If lost small characters, reduce this value to detect more objects.
In manga: text in background texture is busy, should increase this value not to detect background noise.
1.8 08/23/2023
- Fix behavior of deleting item on result view
- Fix crash bug on the older devices. Add option CPU fallback option.
1.7 08/22/2023
- Add import images from pdf file.
- Add export OCR results as searchable pdf file.
- Fix transformer process to reduce OCR error.
- Add edge crop of image to process OCR.
- Add rotate 90 degrees direction and auto rotate detection.
more Version 3.15 08/23/2025
Data Not Collected The developer does not collect any data from this app.
