Skip to content

Windows OCR language support

Windows Built-in OCR uses the OCR language support installed on your device. Use this guide when the language you need is missing in Float Translate or when recognition becomes inaccurate after the text language changes.

Windows exposes only the OCR languages whose support is installed. The selected OCR language should normally match the text you want to recognize.

For example, if only Chinese OCR support is installed and selected, Windows may still return some English letters or words, but English accuracy and word boundaries are not guaranteed. Install English OCR support and select English when you need reliable English recognition.

This applies whenever Windows Built-in OCR processes screen pixels, including Screenshot OCR, Screen Text Capture fallback, region translation, and workflows that select this OCR capability.

Windows labels can vary slightly by version, but the usual path is:

  1. Open Windows Settings → Time & language → Language & region.
  2. Add the language you need, or open Language options from the menu beside an installed language.
  3. Install the OCR component offered by Windows. You do not need to make that language your Windows display language.
  4. Return to Float Translate and select Windows Built-in OCR, then choose the matching recognition language.

If a newly installed language does not appear immediately, close and reopen Float Translate so it can read the current Windows OCR language list again.

Windows OCR can use only the OCR language support installed on the device. For current Windows details, see Available OCR languages and PowerToys Text Extractor: OCR language packs.

If OCR is missing or will not install in Settings

Section titled “If OCR is missing or will not install in Settings”

If Language options does not show OCR, installation does not finish, or Float Translate still cannot find the language afterward, you can install the corresponding Windows OCR capability directly.

Open Windows PowerShell as administrator and list the OCR languages and states provided by your version of Windows:

Terminal window
Get-WindowsCapability -Online | Where-Object { $_.Name -Like 'Language.OCR*' }

The capability name contains the language code. For example, en-US is English (United States), zh-CN is Simplified Chinese, zh-TW is Traditional Chinese, ja-JP is Japanese, and ko-KR is Korean. Use the actual command output as the source of truth. If the target language is absent from the list, Windows OCR does not support it; choose another OCR capability in Float Translate.

After finding the target language, install it by its language code. If you mainly recognize English, install an English OCR language; the example below uses English (United States), en-US. English OCR may detect some text written with the Latin alphabet, but it does not replace the matching OCR support for French, German, Spanish, or other languages. For reliable recognition, install and select the language of the source text.

Terminal window
Get-WindowsCapability -Online | Where-Object { $_.Name -Like 'Language.OCR*en-US*' } | Add-WindowsCapability -Online

Installation can take several minutes. Run the query again afterward and check the target language state:

  • Installed: installation is complete. Fully quit and reopen Float Translate so it can reload the language list.
  • NotPresent: Windows supports the language, but it is not installed.
  • InstallPending: Windows Update or a restart must finish first. Complete Windows Update and restart the device instead of repeatedly running the install command.

If the installation command fails, make sure Windows Update can download optional features. Devices managed by a company or school may be restricted by the organization network, WSUS, or Group Policy, in which case you may need to contact the device administrator. Offline devices may require a Languages and Optional Features source that matches the installed Windows version. See Microsoft’s guidance for acquiring optional Windows content and avoid system packages or registry workarounds from untrusted sources.

  • For ordinary OCR, open Settings → OCR, select Windows Built-in OCR, and choose the recognition language shown below it.
  • In the OCR panel, you can change the OCR capability or language before choosing Recognize again.
  • In a workflow, open the OCR step and select the OCR capability and language that match the source text.

Choosing Use system language follows the Windows language preference available to the built-in OCR engine. Select a specific installed language when the source text differs from that preference.

  • Make sure the selected language matches the text rather than only the Windows display language.
  • Use a sharper image, increase the text size, and avoid busy or low-contrast backgrounds.
  • For mixed-language text, try the language that covers the most important part of the result.
  • Choose another OCR capability in the OCR panel or workflow when Windows Built-in OCR is not suitable for the content.