I keep finding text in places where text is not supposed to be editable. A tracking number arrives in a screenshot. An error message is trapped in a photo from someone else's phone. A scanned PDF looks like a document but behaves like a stack of pictures.
The usual response is to put the image beside the destination and retype it. That is tolerable for six characters. It is miserable for a paragraph, and risky for anything where one wrong digit matters. This is exactly what optical character recognition is for, but I wanted the useful part to feel like copying, not like starting a document conversion project.
Start with the pixels you already have
If an image is already in Droppy's Clipboard, open it and choose Extract Text. The same action is available for an image or PDF sitting in the Shelf or Basket. For something still on screen, Droppy's screenshot capture can select the area first and run OCR on the result.
That matters because the best starting point changes. Sometimes I already copied the screenshot. Sometimes a PDF is waiting with the rest of a project. Sometimes I only need two lines from a web page, video frame, or dialog and do not want to save a screenshot at all. The extraction step should meet the material where it is.
What Droppy actually does with the image
Droppy uses Apple's Vision framework for recognition. It asks for accurate recognition, enables language correction, and lets Vision detect the language automatically. The languages set on your Mac are also used as preferred hints when Vision supports them. Recognition runs on the Mac, so the image does not need to be uploaded to a separate OCR service.
The resulting lines are ordered for reading before Droppy joins them into plain text. OCR cannot infer every bit of layout from every image, but preserving a sensible top-to-bottom order makes the result much more useful when you paste it into Notes, Mail, a support reply, or a search field.
For text still on screen, capture the area and let Droppy recognize it.
- Start Droppy's screenshot capture.
- Select the area that holds the text.
- Droppy runs OCR and shows the recognized text.
An image you already have can go straight through recognition.
- Open the image in the Clipboard, or drop it into the Shelf or Basket.
- Choose Extract Text.
- Read the result, or paste it when Auto-Copy is on.
PDFs get a little more care than a plain image.
- Drop the PDF into the Shelf or Basket.
- Choose Extract Text.
- Pages with real text stay text; scanned pages are recognized.
Recognition runs on your Mac, so the image is never uploaded.
Images and scanned PDFs take slightly different paths
For an image, Droppy sends the pixels through text recognition. PDFs get a little more care. If a page already contains embedded selectable text, Droppy uses it directly. If the page is a scan, Droppy renders that page and recognizes the rendered image. Multi-page results keep page separators, so the text does not become one unexplained wall.
This is why the feature is useful beyond screenshots. Old scanned paperwork, exported receipts, and image-only PDFs can all end up in the same plain-text workflow.
Turn on OCR Auto-Copy when the result window is one step too many
By default, Droppy shows the recognized result so you can read it before copying. I like that when the image is noisy or the text is important enough to check.
For the everyday case, open Droppy Settings, search for OCR Auto-Copy, and enable Auto-Copy OCR Text. After that, a successful extraction skips the result window and copies the recognized text instantly. Droppy only takes that shortcut when the result contains visible text. An empty recognition result still opens the result window instead of quietly replacing your clipboard with nothing.
That small condition is important. Auto-Copy should remove a confirmation step, not remove feedback when recognition had nothing useful to return.
A few ways I use it
Copy an error without retyping it
Capture the dialog, extract the exact wording, then paste it into an issue or search. Error messages are particularly unforgiving of creative transcription, so copying the recognized result is much better than typing what you remember.
Pull a reference from a screenshot
Order numbers, addresses, serial numbers, and tracking codes are often shared as images. OCR turns them back into text you can search or paste. I still check important numbers against the image, because recognition is not magic, but checking is faster than typing from scratch.
Recover text from a scanned document
Drop the PDF into the Shelf or Basket and use Extract Text. Pages with real text stay real text; scanned pages go through recognition. You get one result that is much easier to quote, search, or archive.
OCR is a copy tool, not a truth machine
Clear, upright, high-contrast text gives recognition the best chance. Tiny type, decorative fonts, motion blur, and a photograph taken at an angle can all introduce mistakes. For a casual sentence that may not matter. For an account number or command, compare the result with the source before using it.
I think that is the honest place for OCR on a Mac. It removes the dull work of retyping and gives you an editable starting point. It does not ask you to pretend the starting point is infallible.
From screenshot to paste
The feature is deliberately small. Find the image, choose Extract Text, and paste. If you do it often, enable OCR Auto-Copy and remove the middle window. The recognition stays on your Mac, the output is ordinary text, and the screenshot can go back to being a reference instead of a typing assignment.