Skip to content
ocrogramocrogram

Apple Live Text limitations (and how to automate OCR)

Live Text needs a click and a drag every time. Here is what it will not do, and how to copy screenshot text automatically.

Live Text is the right default. It is fast, private, and already on the Mac. People look for another tool when they notice it is a gesture, not a pipeline.

What Live Text will not do

  • It will not copy text when you take a screenshot. You still have to open the image (or click the thumbnail) and select.
  • It will not run in the background. There is no login item, no folder watch, no "always OCR screenshots" switch.
  • It will not expose a first-party CLI. Shortcuts can get you partway there; a shell script cannot call "Live Text" the way it can call tesseract.
  • It will not help if you never look at the screenshot. The text stays in the pixels until you go get it.

None of that is a bug. Live Text is a pointer tool. The gap is automation.

Automate it with ocrogram

ocrogram uses the same Apple Vision API Live Text uses. The difference is the trigger: a new file in your screenshot folder, not a click. Take Cmd+Shift+3 or 4. Wait a beat. Cmd+V.

brew install joelpeckham/ocrogram/ocrogram
ocrogram start

Recognition stays on-device. Install, requirements, and troubleshooting.

When to keep using Live Text

Photos you did not just capture — a camera roll, a downloaded PNG, a page in Safari — still belong to Live Text. ocrogram only watches the screenshot folder. For a one-off image, open it and drag. For the tenth error dialog of the afternoon, let the daemon have it.

Related