Instantly extract text and LaTeX math from images and documents on your Mac. State-of-the-art OCR performance powered by AI, tailored for each use case.

Check prices

Features

FrameOCR general settings UI. FrameOCR CJK vertical text UI.

LaTeX Math Recognition in Action:

Release Notes

Release Version: v1.2.3

Date: 2025-10-07

Version Description:

Fixed bug where the text preview would show up in the wrong place when multiple monitors are connected.

Release Version: v1.2.2

Date: 2025-09-21

Version Description:
  • Added support for Bulgarian.
  • Added tooltip pop-ups for each setting when hovering the mouse over it.

Release Version: v1.2.1

Date: 2025-09-17

Version Description:
  • Added support for Japanese, Chinese, and Korean vertical text.
  • Added support for Vietnamese and Thai.
  • Added 'About' section in settings.
  • Added a button under General Settings to reset all settings and keyboard shortcuts. Does not affect license settings
  • Further improvements for LaTeX processing.
  • Fixed a bug where running OCR multiple times too quickly would cause the app to crash.

Release Version: v1.1.1

Date: 2025-09-14

Version Description:
  • Added support for Russian, Belarusian, Ukrainian, and Korean.
  • Added a small pop-up notification window that shows scanned text after running OCR (goes away after 5 seconds, like normal macOS notifications).
  • Improved processing for LaTeX operators (sup, inf, Pr, log, ln, lim, det, dim, exp, max, min, sin, cos, tan, csc, sec, cot).
  • Updated the app + menubar icons.

Release Version: v1.0.0

Date: 2025-09-10

Version Description:

Initial public release of FrameOCR.

Frequently Asked Questions (FAQ)

What versions of macOS are compatible?

FrameOCR is compatible with macOS Sonoma (14.0) and up.

Is there a free trial?

FrameOCR is free to use and evaluate, and after 7 days it will start asking you to consider upgrading. If it helped you be more productive and saved you time, please consider purchasing the app!

Is the purchase a one-time payment or a subscription?

Your purchase is a one-time payment, not a subscription.

What languages are supported by the text recognition engine?

English, Afrikaans, Albanian, Belarusian, Bosnian, Bulgarian, Chinese, Croatian, Czech, Danish, Dutch, Estonian, French, German, Hungarian, Korean, Icelandic, Indonesian, Irish, Italian, Japanese, Latin, Lithuanian, Malay, Maori, Norwegian, Occitan, Polish, Portuguese, Russian, Serbian, Slovak, Slovenian, Spanish, Swahili, Swedish, Tagalog, Thai, Turkish, Ukrainian, Uzbek, Vietnamese, Welsh.

The Chinese OCR processing supports Simplified Chinese (简中), Traditional Chinese (繁中), and Pinyin (拼音).

For Chinese, Japanese, and Korean, auto-detection of text orientation is also supported (as of v1.2.0). If the auto-detection isn't working (which may happen for strangely structured text), you can also manually specify this setting to horizontal or vertical. If the detected text is vertical, then a right-to-left reading direction is assumed.

Remember to select your target language in the app settings so that the best model can be used (Settings > General > OCR Language).

Does FrameOCR use cloud OCR processing?

No, FrameOCR runs all OCR processing entirely locally and offline on your machine. That means there is no limit to how much you can use the text or LaTeX recognition features.

What programming language is FrameOCR written in?

FrameOCR is written entirely in Swift 6, built natively for macOS.