How to Watch YouTube with Live Translated Captions

YouTube can auto-translate captions — but only when a video has captions, and live streams and many smaller channels often don't. To get translated captions for anything playing in a tab, open the AI Live Voice Translator in Chrome or Edge, choose Tab or screen sound, start listening, and share the YouTube tab with its audio. The speech is recognised and translated into your language as the video plays.

Set it up step by step

1. Open the YouTube video or live stream in a Chrome or Edge tab.
2. In another tab, open the AI Live Voice Translator.
3. Set Listen to to Tab or screen sound.
4. Choose the video's Spoken language, or Auto-detect.
5. Choose Translate into — your language.
6. Click Start listening, pick the YouTube tab in the sharing dialog, and tick Share tab audio.
7. Play the video. Captions and translations appear phrase by phrase.

Click Floating captions for a small caption window that stays on top of the video (Chrome and Edge on a computer).

YouTube's built-in option first

For videos that already have subtitles:

1. Click the CC button to turn captions on.
2. Open Settings (the gear) › Subtitles/CC › Auto-translate, and choose your language.

That's quick when it's available. The live translator is for everything else: live streams, videos without captions, captions in the wrong language, and audio from podcasts, webinars and online courses on other sites.

Works on more than YouTube

The same method works for any site that plays sound in a browser tab:

• Live streams on YouTube, Twitch and other platforms.
• Online courses, lectures and webinars.
• News broadcasts and podcasts played in the browser.
• Social media videos on desktop.

Privacy and engines

Speech from a shared tab is recognised in your browser by OpenAI's Whisper model, so the audio never leaves your computer. For translation, choose:

• Gemini AI — the most natural translations; the recognised text (not audio) is sent to iPixHub's server, not stored, and metered daily.
• Chrome built-in — on-device and unlimited in desktop Chrome 138 and later.

The Whisper model downloads once and is cached. On an ordinary laptop, expect captions a few seconds behind the speech; a computer with a graphics card (WebGPU) is faster.

Tips for the best captions

• Clear speech works best: interviews, lectures, news and vlogs caption well; music videos and noisy scenes don't.
• Pause or slow down if the speaker is fast — YouTube's playback speed setting (0.75×) gives the captions time to keep up.
• Choose the spoken language instead of auto-detect when you know it — recognition is more accurate.
• Copy the transcript after watching to keep notes, or translate a cleaned-up version with the AI Translator.

Want the audio itself?

To save the sound of a video file you own — a lecture recording, a webinar you downloaded — as an MP3 for listening offline, use Extract Audio. Respect creators' rights: only download what you're allowed to.

Use it as a learning aid, not a source of truth

Live speech recognition and machine translation are both usually good, and occasionally confidently wrong. They're great for following content in a language you don't speak, but don't quote them as exact or rely on them for important decisions.

Frequently asked questions

Does it work on a phone?

Sharing a tab's audio needs a desktop browser. On phones, YouTube's own auto-translate is the practical option.

Is anything recorded?

No. The captions are generated live and the audio isn't saved.

Why are the captions behind the video?

Each phrase is translated when the speaker pauses, and recognition takes a moment on slower computers. Slowing the video helps.