Extract text from a screenshot, a video frame or a scanned page without retyping any of it. Press ⌘⇧T, drag over what you can see and cannot select, and it lands on your clipboard as editable text.
Free to use. Recordings are capped at three minutes and export at 720p with a watermark.

An error message that has to go into a ticket word for word. A number on a slide that arrived as a picture. The characters are right there on the display, and there is nothing to drag a cursor through.
Retyping is where the typos come from, and a wrong character in a licence key or a stack trace is its own problem to find. Turning a screenshot into text takes a keystroke instead: Capture Text puts a box over the region and hands the characters back, already on the clipboard.
A shot you are marking up does not need a second pass. Extract Text sits in the editor’s own menu, next to Adjustments and Beautify, and reads whatever part of the image you draw a box around.

The exact wording of a crash, into a ticket, without transcribing it by hand and getting one character wrong.
A command or a URL that only exists inside a video, where pausing still leaves you nothing to select.
A receipt, a contract or a form that arrived as a picture of a page rather than as a document.
The figures off something a colleague pasted in as an image instead of sending the table it came from.
A PDF lands where your captures land, with its pages down one side. It is the same editor, opened on a document instead of a screenshot.

Text extraction is free with a monthly allowance, and unlimited on Pro and Studio.
Drag over part of a window, a menu, a video or another app. If it is on the display, it can be read.
The text is copied the moment the scan finishes, and kept in a history you can go back to.
Five extractions per calendar month, reset on the first. Pro at $39 once and Studio at $99 once make it unlimited.
A capture you are already annotating has Extract Text in its menu, over any selection you draw.
A PDF page with no text layer behind it is just pixels. The same scan reads it, a region at a time.
The script is detected from the region rather than assumed, so text that is not English comes back as it was written.
The table you just pulled the addresses out of is still open, with the blur tool on the same toolbar. Anything in the shot that was never meant to travel with it gets covered before the image goes anywhere.

The OCR runs on your Mac. The region is not uploaded for processing, so an internal dashboard or a signed contract can be read without a copy of it leaving the machine. You do need to be signed in to use OwnClip, so it is not usable fully offline — the reading itself is local either way.
The free allowance is five extractions per calendar month, shared with the Clean Up, Rephrase and Summarize actions in the text window, and it resets on the first. Pro is $39 once and Studio is $99 once; both make it unlimited, and neither is a subscription.
Press ⌘⇧T and drag a box over the part of the screen holding the text. OwnClip reads the region and puts the result on your clipboard, and keeps a copy in the text window so you can go back to it. The image can be a screenshot, a photo, a video frame or another app’s window.
Five times per calendar month on the free tier, counted on your own Mac and reset on the first of the month. The same pool covers the Clean Up, Rephrase and Summarize actions in the text window. Pro at $39 once and Studio at $99 once both make it unlimited.
Yes. A PDF opens in the same editor as your captures, and a page with no text layer behind it is read the way any other region of the screen is read — drag over the part you need. The document stays on your Mac.
No. Text recognition runs on your own machine and the region is never sent anywhere for processing. Nothing leaves the Mac unless you choose to share it, and sharing uploads to your own Google Drive rather than to our servers.
OwnClip needs macOS 26 or later on an Apple silicon Mac. It is a native Mac app — there is no Windows or web version.
Free includes the whole editor — timeline, camera effects, audio cleanup, captions, annotations, and the screenshot and PDF editors. What it caps is the take itself: three minutes, recorded at 720p, with a small watermark composited into the frames as they are written. You can keep 25 recordings. A few tools are paid rather than capped — Smart Zoom and Scroll Capture are Pro; Brand Studio, Go Live and Step Movie are Studio. Pro is $39 once and lifts every cap. Text extraction has its own allowance of five uses a month.
Free to start. macOS 26 or later, Apple silicon.