Follow spoken audio with Live captions
Turn on Live captions and check language and microphone options while accounting for transcription errors.
Publication & corrections: infostoa · Editorial standards
Individual author and specialist reviewer information is not provided for this article.
Setup and requirements
Press Windows+Ctrl+L or open Settings > Accessibility > Captions and enable Live captions. On first use, review the on-device speech-processing prompt and download a caption language. Live captions requires Windows 11 version 22H2 or later; translation has narrower device and version requirements. In the caption window’s settings, dock it above or below the desktop or float it over the screen so it does not cover the video or app controls.
A practical use
Use captions when following system audio in a meeting, lecture, or video is easier with text. Adjust position and style for readability. The option to include microphone audio can caption your own speech, but it requires microphone permission; overlapping speakers can still cause gaps. Names, technical terms, and numbers are especially error-prone, so compare important passages with the original audio or an official meeting record.
Practitioner perspective
Microsoft Support documents on-device caption processing and the Windows 11 version requirement. Windows Central’s Mauro Huculak describes captions as a built-in way to read audio across apps. Real-time recognition remains fallible: resource-heavy screen sharing or overlapping speech can delay or drop captions. Do not use the output as the sole record for legal proceedings, clinical decisions, or precise meeting minutes; pair it with a verified transcript or source recording when accuracy matters.
Limits and troubleshooting
To add a caption language, use the caption window’s language settings and install the speech component Windows requests. Translation has additional limits, including Copilot+ PC and Windows 11 version 24H2 or later; a PC that supports captions may not show translation. If text lags or drops, reduce resource-heavy video effects and check the microphone and permission if microphone audio is included. Test the selected language before a live session because recognition quality changes with language and audio conditions.