Published on

Copy, OCR, and Paste

Authors
  • avatar
    Name
    Brian Weeks
    Twitter

Copy, OCR, and Paste

I often need to grab text from screenshots, slides, or video frames. Retyping it manually is a pain. Yes, you could take a screenshot, open the image in Preview, go to Tools->Text Selection, copy your desired text and paste it. But, that's too slow and cumbersome.

So I built a workflow that lets me:

  1. Take a screenshot of some text with Shift+Command+4
  2. Press a custom keyboard shortcut Shift+Command+O
  3. Paste plain text anywhere!

It’s fast and entirely offline—powered by Tesseract OCR and AppleScript.

How it Works

  1. Detects if your clipboard contains an image.
  2. Saves it temporarily as a file.
  3. Sends the file to Tesseract OCR.
  4. Places the OCR results back in your clipboard.

How to Set It Up

1. Install Tesseract OCR

First, install Tesseract using Homebrew: brew install tesseract

Take note of where tesseract has been installed, as you'll need to reference the path in the AppleScript. /opt/homebrew/bin/tesseract

2. Save the AppleScript as a Quick Action

  1. Open Automator.
  2. Create a new Quick Action.
  3. In the "Workflow receives current" dropdown, choose: no input in any application.
  4. Drag in a "Run AppleScript" action.
  5. Paste the full script (below) into the AppleScript block.
  6. Save it with a recognizable name like "Clipboard Image OCR".

3. Assign a Keyboard Shortcut

  1. Open System Settings → Keyboard → Keyboard Shortcuts → Services (or Quick Actions).
  2. Scroll to find your "Clipboard Image OCR" quick action.
  3. Add a custom shortcut, like ⌘⇧O (or whatever you prefer).

📜 The AppleScript Code

on run
    -- Hardcoded paths:
    set tesseractPath to "/opt/homebrew/bin/tesseract"
    set languageCode to "eng" -- Tesseract language code
    
    set userHome to POSIX path of (path to home folder)
    set tempDir to userHome & "Downloads/clipboard_temp/"
    
    try
        --------------------------------------------------------------------------------
        -- 1. Check if there is an *image* in the clipboard
        --------------------------------------------------------------------------------
        set clipboardInfo to clipboard info
        set clipboardInfoText to (clipboardInfo as text)
        
        if (clipboardInfoText contains "TIFF picture") ¬
            or (clipboardInfoText contains "JPEG picture") ¬
            or (clipboardInfoText contains "GIF picture") ¬
            or (clipboardInfoText contains "PNGf") then
            
            --------------------------------------------------------------------------------
            -- 2. Attempt to read the clipboard as PNG first, if that fails, try TIFF, etc.
            --------------------------------------------------------------------------------
            set imageData to ""
            set pickedFormat to ""
            
            -- Try PNG
            try
                set imageData to the clipboard as «class PNGf»
                set pickedFormat to "PNG"
            on error
                -- If PNG fails, try TIFF
                try
                    set imageData to the clipboard as TIFF picture
                    set pickedFormat to "TIFF"
                on error
                    -- If TIFF fails, try JPEG
                    try
                        set imageData to the clipboard as JPEG picture
                        set pickedFormat to "JPEG"
                    on error
                        -- If JPEG fails, bail out
                        error "Clipboard image could not be read as PNG/TIFF/JPEG."
                    end try
                end try
            end try
            
            --------------------------------------------------------------------------------
            -- 3. Write the raw image data to a temporary file
            --------------------------------------------------------------------------------
            do shell script "mkdir -p " & quoted form of tempDir
            
            -- Decide on an extension based on the format we picked
            if pickedFormat = "PNG" then
                set tempImagePath to tempDir & "clipboard_image.png"
            else if pickedFormat = "TIFF" then
                set tempImagePath to tempDir & "clipboard_image.tiff"
            else if pickedFormat = "JPEG" then
                set tempImagePath to tempDir & "clipboard_image.jpg"
            else
                -- fallback
                set tempImagePath to tempDir & "clipboard_image.bin"
            end if
            
            set fileRef to (open for access (POSIX file tempImagePath) with write permission)
            set eof fileRef to 0
            write imageData to fileRef
            close access fileRef
            
            --------------------------------------------------------------------------------
            -- 4. Run Tesseract on the saved file
            --------------------------------------------------------------------------------
            set ocrOutputPrefix to tempDir & "ocr_output"
            set ocrOutputFile to ocrOutputPrefix & ".txt"
            
            set tesseractCommand to quoted form of tesseractPath & space & ¬
                quoted form of tempImagePath & space & ¬
                quoted form of ocrOutputPrefix & " -l " & languageCode
            
            do shell script tesseractCommand
            
            --------------------------------------------------------------------------------
            -- 5. Read the OCR result, remove line breaks, copy to clipboard, then clean up
            --------------------------------------------------------------------------------
            -- Original:
            -- set ocrResult to do shell script "cat " & quoted form of ocrOutputFile
            -- set the clipboard to ocrResult

            -- Revised (removes all '\r' and '\n'):
            set ocrResult to do shell script "cat " & quoted form of ocrOutputFile & " | tr -d '\\r\\n'"
            set the clipboard to ocrResult
            
            do shell script "rm -rf " & quoted form of tempDir
            
            -- Return value (no dialog)
            return "Clipboard image converted to OCR text (line breaks removed)."
            
        else
            -- Return value (no dialog)
            return "No image found in clipboard (Info: " & clipboardInfoText & ")"
        end if
        
    on error errMsg
        -- Cleanup in case of error
        try
            do shell script "rm -rf " & quoted form of tempDir
        end try
        
        -- Return value (no dialog)
        return "Error: " & errMsg
    end try
end run