ScanLingo: Learn by ReadingView on the App Store

ScanLingo: Learn by Reading · Reading learning guides

How to turn photo object candidates into vocabulary cards

Updated 2026-08-29

Treat general-object candidates as prompts, then check and edit them before saving local vocabulary cards.

  1. Capture and check the original

    Object vocabulary uses the bundled OpenMMLab RTMDet-s FP16 Core ML model on the device to generate general-object candidates from an image you choose or take; you may review, edit, and save vocabulary. The model is not downloaded at runtime, and the current release does not send source images to the operator's servers. The bundled model file is approximately 20.2 MB.

    Camera and photo access are requested only when you choose the related feature, and you can change access at any time in iOS Settings.

  2. Use language tools to clarify meaning

    Machine recognition, translation, and speech output may be incomplete or inaccurate and must not replace human review in professional or high-risk contexts.

    Lighting, clarity, layout, and language affect OCR accuracy. Check the result against the original before relying on it.

  3. Keep a route back to the source

    Bind recognised material to a book and page number, organise it as sentence cards, and return to that context while reviewing local learning progress and reports.

    Book and page associations, sentence cards, learning profiles, and reports are stored on your device by default. The current release has no remote account or sync feature.

  4. Understand accuracy and local-storage boundaries

    Images, OCR text, translations, and learning content are processed and stored on the device by default. Speech-input results are stored on the device by default; the operator does not retain microphone audio.

    The current release has no advertising and no analytics. It also does not include advertising tracking, cross-app tracking, IDFA use, heat maps, or session replay.

Learn More

Photo translation and contextual review: FAQ

These answers describe the current ScanLingo workflow for personal reading and language study.

What reading material is photo translation suitable for?
Use the camera or choose an image from the system photo library. You can review the OCR result before translating or organizing it.
Can I correct text after image OCR?
Lighting, clarity, layout, and language affect OCR accuracy. Check the result against the original before relying on it.
Why connect sentence cards to a book and page?
Bind recognised material to a book and page number, organise it as sentence cards, and return to that context while reviewing local learning progress and reports.