Why the cursor can't help here

macOS only lets you copy text that an application has marked as selectable. Images, rendered PDFs, paused video frames, terminal screenshots, virtual machine windows, and many error dialogs are not selectable even when the letters look perfectly crisp. Re-typing is slow, error-prone, and the fastest way to mistype a serial number, a hex hash, or a URL.

The right tool for this job is on-device OCR: an app that reads the pixels in a region you select, recognizes the text locally, and hands you a clean string you can paste anywhere. Saither builds one called PixelScan. It uses Apple Vision, so recognition runs on your Mac and no image is sent to a server.

What you need. PixelScan runs on macOS 14 or later on Apple Silicon (M1 or later). The first time you trigger a capture, macOS will ask you to grant Screen Recording in System Settings → Privacy & Security.

Step-by-step workflow

  1. Install and launch PixelScan. Open the DMG, drag the app to Applications, and launch it. The app lives in the menu bar — there is no main window to manage.
  2. Grant Screen Recording when prompted. The first capture triggers the standard macOS prompt. Open System Settings → Privacy & Security → Screen Recording, toggle PixelScan on, and relaunch the app. See the permissions reference for why this is required.
  3. Start a capture. Use PixelScan's configured capture shortcut to bring up the region selector. Move the cursor to one corner of the text you want, click and drag to the opposite corner, and release.
  4. Review the recognition result. PixelScan uses Apple Vision locally and returns clean text quickly for typical regions.
  5. Copy and paste. Copy the recognized text, switch to the app that needs it, and paste with ⌘V. PixelScan also keeps the capture in its searchable scan history, so you can return to it later.

Where this works (and where it doesn't)

Screen OCR is most useful when text is visible but blocked from normal selection. The following are reliable targets:

  • Error dialogs and system messages rendered as images.
  • Subtitles, captions, and on-screen identifiers inside paused video.
  • Slides, lecture notes, and reference images.
  • PDFs shown in viewers that block text selection.
  • Remote desktop and virtual machine windows.
  • Serial numbers, addresses, and labels on photographed screens.

It is not designed for bulk document processing, layout-preserving PDF conversion, or handwriting recognition. For dense code with unusual symbols, always double-check the result against the source — OCR accuracy on individual characters is high but never 100%.

Permissions and privacy, plainly

PixelScan needs Screen Recording because macOS only allows an app to read pixels from a region of your screen after that grant is in place. PixelScan uses those pixels solely to run the OCR you requested. The captured image is not uploaded — recognition happens through Apple Vision on your Mac, and no network request is made for text recognition. The full list of permissions used by Saither apps is on the permissions page.

Pair it with a clipboard history

Once you start capturing text from the screen regularly, you'll want to keep what you've copied. Pair PixelScan with ClipDeck, which records every text copy in a searchable list with pin support, so a captured serial number or URL is never lost. ClipDeck needs no special permission for its current features.

Related guides

PixelScan — on-device screen OCR for Mac

Draw a region, get clean text. Apple Vision, $5.99 one-time, no subscription.

Common questions

Why does PixelScan need Screen Recording permission?

macOS requires Screen Recording before any app can read pixels from a region of your screen. PixelScan uses those pixels only to perform the capture you request.

Does the captured image leave my Mac?

No. Text recognition runs on-device through Apple Vision. No network request is made for the OCR step, and no third-party service receives the image.

What can I capture text from?

Anything visibly displayed on screen: images, video frames, PDFs, application dialogs, remote desktop windows, virtual machines, and web pages.

What happens if I deny Screen Recording?

PixelScan's capture command will not be able to read the pixels in the region you select. You can grant the permission later in System Settings and the feature will work immediately.

What Macs are supported?

macOS 14 or later on Apple Silicon (M1 or later). Intel Macs are not supported.

At a glance

Permission
Screen Recording (capture only). See permissions reference.
Network
None for OCR. Recognition runs through Apple Vision on-device.
Platform
macOS 14+, Apple Silicon (M1 or later). Intel not supported.
Price
$5.99 USD one-time. Free 3-day trial. Buy or download.

Need more help? Email [email protected] or visit the support page.