Screenshot Translation

Screenshot Translation helps with visible text that cannot be selected directly, such as video subtitles, images, games, the visible page of a scanned PDF, and remote desktops. It works as one continuous action: after you select a screen region, Float Translate recognizes the text, brings it into the translation panel, and starts translating it.

Screenshot Translation uses an OCR capability and the translation capabilities available for the current language pair. You do not need to build a workflow before using it.

Set up Screenshot Translation

In the main panel, open Shortcuts under Settings. In the Translation card, assign a global shortcut to Screenshot Translation.

On macOS, screenshot recognition needs permission to read screen pixels. If Float Translate asks for access when you start the task, follow the in-app prompt. You can also open Permissions under Settings in the main panel to review the current status.

Translate visible text on the screen

Make sure the content you want to translate is visible on the screen.
Press the Screenshot Translation shortcut.
Drag to select the region that contains the text.
Release the pointer and wait for recognition to finish.
The translation panel opens with the recognized text and starts translating it.
Check the recognized source and the translations. If recognition is incorrect, edit the source text and translate it again.

Improve recognition

Select only the text you need. Avoid including unrelated icons, borders, or other content.
Use text that is clear and large enough to read. Heavy blur, obstruction, or rotation can reduce recognition quality.
Keep multiline text in its normal reading order. Do not include widely separated regions in one capture.
If only a small part of the source is wrong, edit it in the translation panel instead of capturing the screen again.

Review and compare results

Recognized text opens in the standard translation panel. You can change the source or target language, edit the recognized source, compare results from multiple translation capabilities, and copy the translation you want to use.

Choose Screenshot Translation or Screenshot OCR

Use Screenshot Translation when your goal is to understand visible text as quickly as possible.

Use Screenshot OCR when you need to keep the captured image, inspect recognition line by line, organize the text, or copy the original without translating it. If the text can already be selected, Translate Selected Text is usually faster and avoids OCR recognition errors.

If the task does not finish

The capture cannot start on macOS: Open Permissions in the main panel, review Screen Recording access, then retry the task.
A protected video or app cannot be captured: Some content prevents system screenshots. Float Translate cannot bypass that restriction.
Recognition does not start: Open Services and confirm that an OCR capability is enabled, configured, and currently available.
No text is recognized: Select a smaller region and retry with clearer text. If you use Windows built-in OCR, also confirm that the Windows OCR language required by the source text is installed and selected.
Text is recognized but no translation appears: Open Services and confirm that at least one translation capability is enabled, configured, and available for the current language pair.