You can see the text. You need the text. But copying does not work. Whether it is a video, a popup, or a protected PDF -- on macOS, there is now a solution.
A Problem Everyone Knows
You are watching a tutorial on YouTube. The speaker shows a configuration, a command, a URL. You want to copy the text -- but there is nothing to select. It is a video. The text is part of the image.
Or you are working with a web form. An error message appears as a popup -- and disappears again after three seconds. You did not read fast enough. The text is gone.
Or you open a PDF that someone exported from another system. The text is visible, but selection is disabled. Copy? Not possible. "Protected."
In all these cases, you probably do the same thing: you type the text out manually. Character by character. And wonder why this is still necessary in 2026.
Freeze the Screen -- Copy the Text
The idea behind FreezeText is simple: if you can see text on your screen, you should be able to copy it. Always.
FreezeText freezes the entire screen at the press of a key. Everything stands still -- videos, animations, popups. You then draw a rectangle around the text you need. The integrated text recognition (OCR) via Apple's Vision framework identifies the characters and places the result directly in your clipboard.
No upload, no external service, no waiting. Processing happens entirely on your Mac.
Where FreezeText Helps
Copying Text from Videos
Tutorials, recorded presentations, webinars -- anywhere text is embedded in moving images. Traditional copying fundamentally does not work here because the text is not a text element but part of the video frame. FreezeText pauses the image and makes the visible text selectable.
This applies beyond YouTube. Screen recordings from colleagues, recorded meetings, or training videos all follow the same mechanism: as long as the video is playing, the text is unreachable.
Disappearing Popups and Tooltips
Error messages, notifications, hover tooltips -- many UI elements are intentionally short-lived. They appear for a few seconds and then vanish. With FreezeText, you freeze the screen before the element disappears. The tooltip stays put, and you can select the text at your leisure.
Particularly helpful with ticketing systems like Jira or Confluence, where hover effects regularly prevent text selection.
Protected PDFs
Some PDF files have copy restrictions. Text selection is disabled even though the text is clearly readable. This frequently happens with official documents, contracts, or reports exported from accounting systems.
Since FreezeText does not extract the PDF text but reads the visible screen content via OCR, it bypasses these restrictions. What you can see, you can copy.
Text from Images and Screenshots
Scanned documents, infographics, error messages shared as screenshots -- anywhere text exists as pixel graphics. FreezeText recognizes the text and returns it as editable plain text.
How It Works
- Press the hotkey -- The default hotkey is ⌃⌥⌘T (Ctrl+Alt+Cmd+T). FreezeText then freezes the entire screen.
- Select an area -- Draw a rectangle with your mouse around the text you want to copy.
- Text is copied -- OCR recognition runs instantly. The recognized text is in your clipboard. Optionally, a text editor popup opens for post-processing.
FreezeText lives in the menu bar and supports multiple displays simultaneously. Smart text processing automatically removes line breaks and fixes hyphenation on demand -- useful when copying text from multi-column PDF layouts.
All captured texts are stored in a searchable history with automatic deletion after a configurable time period.
Privacy: Everything Stays Local
FreezeText exclusively uses Apple's Vision framework for text recognition. Processing takes place directly on your Mac. No data is sent to external servers -- neither the screen content nor the recognized text ever leaves your device.
This is a fundamental difference from online OCR services, where you have to upload screenshots to third-party servers.
Frequently Asked Questions
Does FreezeText work with every video player?
Yes. FreezeText operates at the level of screen content. It does not matter whether you are watching a YouTube video in the browser, a local video in QuickTime, or a Zoom recording. As long as the text is visible on screen, it can be recognized.
How accurate is the text recognition?
Recognition is based on Apple's Vision framework and is very reliable with clearly readable text. With small fonts, low-resolution videos, or unusual typefaces, accuracy may vary. The optional editor popup allows for quick corrections before pasting.
Does FreezeText support multiple languages?
Text recognition supports all languages covered by Apple's Vision framework -- including English, German, French, Spanish, and many more.
Do I need an internet connection?
No. FreezeText works entirely offline. Neither installation nor daily use requires an internet connection.