Add a document
Choose the document to read from the camera, your photo library, or a PDF. Try it with real documents such as vertical text, handwriting, and forms.
Yomikurabe
One document, many OCR engines.
An app for reading images and PDFs and comparing the recognized text side by side.
iOS / Android
Overview
Yomikurabe is a mobile app that reads Japanese documents with several OCR engines and compares the recognized text, processing time, and estimated cost.
You can capture a document with the camera, choose a photo, or pick a page from a PDF. On-device and cloud engines all run on the exact same image, and you can review the results engine by engine.
Yomikurabe is designed with Japanese documents in mind, including handwriting and vertical text.
Features
Compare engines
Select on-device and cloud OCR engines and run them together on the same image. Check the recognized text and recognized regions for each engine.
Keep your results
Save measurements, including processing time and estimated cost, to history so you can review them later or export them as CSV.
Supported engines
On-device
Apple Vision (iOS)
ML Kit Japanese (Android)
RapidOCR (small / medium)
NDLOCR-Lite
PP-OCRv5 server
Cloud
Google Cloud Vision
Gemini
* Cloud engines require an API key or access token for each service. Charges follow the settings of each service.
Sample documents
Printed form
Handwritten document
Vertical text
* These images are sample documents we use to test the app.
Please note
- You can select up to 3 engines at a time.
- When you read with on-device engines, your images are not sent off your device.
Requirements
| iOS: | iOS 18 or later |
|---|---|
| Android: | Android 12 or later |
| Availability: | Available on the App Store and Google Play |
Update History
NOW LOADING ...