Live Caption automatically generates real-time, on-device captions for audio and video playing in the Chrome browser on your Chromebook — YouTube videos, podcasts, video calls, or locally stored media files. It works entirely offline once enabled, since the captioning happens on the device rather than through a cloud service.
How to turn on Live Caption
- At the bottom right of your screen, select the time to open Quick Settings.
- Select Settings, then go to Accessibility.
- Under Audio and captions, select Captions.
- Turn on Live Caption.
Once enabled, captions appear automatically in a small movable box near the bottom of the screen whenever audio or video plays in Chrome — you don’t need to turn it on separately for each video or site.
What to know before relying on it
- Live Caption requires ChromeOS 89 or later.
- It currently supports English only.
- It only captions audio playing through the Chrome browser itself — it won’t caption audio from Android apps or Linux apps running in their own windows.
- If you don’t see the Live Caption option under Accessibility settings at all, your specific device or ChromeOS version doesn’t support it yet.
Moving or resizing the caption box
Once captions appear, you can drag the caption box to a different part of the screen, which is useful if it’s covering video controls or on-screen content you need to see. It stays in that position for future videos until you move it again.
Related accessibility features
Live Caption is one of many built-in accessibility tools on ChromeOS. If you’re setting up a Chromebook for someone who needs broader support — screen magnification, high-contrast mode, or switch access — see our full Chromebook accessibility features guide. If you rely more on speech-to-text for writing rather than captioning what you hear, Voice Typing is the ChromeOS feature built for that instead.
If captions seem inaccurate or lag behind
Accuracy depends on audio clarity — background noise, overlapping speakers, or heavily accented speech all reduce accuracy, since the model runs entirely on-device rather than using a larger cloud-based model. There’s no manual accuracy adjustment; if captions are consistently unusable for your needs, an external captioning or transcription service may work better for that specific use case.
