OCR: Extract Text from Images

What is ShareX OCR?

Optical character recognition (OCR) converts text visible in an image into editable, searchable text. ShareX uses the OCR engine included with Windows to recognize text from a selected screen region or an image file.

Recognition is performed locally by Windows. The result is not sent to an online service unless you choose one of the optional search or translation actions after recognition.

Extract text from the screen

  1. Open Tools -> OCR in ShareX.
  2. Select the screen region containing the text.
  3. Wait for the OCR window to display the recognized result.
  4. Correct the text in the result box if necessary.
  5. Click Copy all to copy it to the clipboard.

The OCR window stays open after recognition. Use Select region to capture another area or Select image to recognize text from an image file without reopening the tool.

You can also create a ShareX hotkey for Tools -> OCR. When the OCR task is launched with an image file path, ShareX recognizes that file instead of asking for a screen region.

Recognition options

Language

Choose the language used by the text in the image. ShareX lists the OCR recognizer languages currently installed in Windows. Selecting the correct language improves word recognition, punctuation, and reading order.

Scale factor

The scale factor enlarges the source before recognition. Increasing it can improve small or low-resolution text, but also takes more processing time and memory. Start with the default value and raise it when small characters are being missed.

Single line

Enable Single line when the result should be joined into one line instead of preserving recognized line breaks. This is useful for serial numbers, commands, addresses, and short interface labels.

Copy recognized text automatically

Enable this option to copy every successful result to the clipboard as soon as recognition finishes. Leave it disabled when you normally need to review or correct the result first.

Changing the language, scale factor, or single-line option reruns recognition on the current source image.

Search, translate, or process recognized text

The Services section can open the recognized text in a web service. The default choices include Google Translate, Google Search, Google Images, Bing, DuckDuckGo, and DeepL.

Choose a service and click Open service. ShareX URL-encodes the current result and inserts it into the service URL. Enable Close after opening if the OCR window should close after the browser opens.

Use the settings button beside the service list to add, edit, remove, or reset services. A custom service URL must contain {0} where the encoded recognized text belongs. For example:

https://example.com/search?q={0}

Opening a service sends the recognized text to that website. Review sensitive text before using an online search or translation service.

OCR task settings and silent mode

Open Task settings -> Capture -> OCR to configure the default recognition language and OCR workflow.

Silent mode is useful for a dedicated OCR hotkey: select a region and receive its text on the clipboard without an additional confirmation window.

When OCR is run on an image file path, ShareX can also write the recognized result to a .txt file beside that image.

Install another OCR language in Windows

ShareX can only use languages exposed by the Windows OCR engine. To install another language in Windows 11:

  1. Open Start -> Settings -> Time & language -> Language & region.
  2. Select Add a language next to Preferred languages.
  3. Find the language, select it, and click Next.
  4. Make sure the language pack is selected and install it.
  5. Wait for Windows to finish installing its language features.
  6. Close and reopen the ShareX OCR window to refresh the language list.

Not every Windows language includes OCR support. If an installed language is not listed in ShareX, check whether that language provides the Windows OCR feature. Microsoft maintains current Windows language installation instructions.

ShareX OCR requires Windows 10 version 1903 (build 18362) or newer.

Improve OCR accuracy

Troubleshooting

The language list is empty or missing a language

Install the language pack in Windows, confirm that the language supports OCR, and reopen the OCR window. Windows editions restricted to a single display language can also limit language installation.

Small text is recognized incorrectly

Select a tighter region and increase the scale factor. If possible, capture the source at its native size instead of recognizing a scaled or compressed screenshot.

Text is copied as one long line

Disable Single line and let ShareX preserve the line breaks returned by Windows OCR.

For AI-assisted image descriptions and broader visual analysis, see the Analyze Image documentation.