A video gives you one subtitle line. This adds the two a learner needs: Pinyin under every Chinese character and a translation on a second line — side by side with the original. Free on YouTube, Netflix and any site with subtitles, including subtitles printed into the picture. No account, no API key. Tap any word for its meaning.
Free on YouTube, Netflix & sites with subtitles — including burned-in subtitles, read on your own computer · No sound-only? Your own Soniox key (~12¢/hr)
1) Free — the video's own subtitles. No key, no cost. It reads the subtitles the video already has and rebuilds them as a learning view: the original line, Pinyin stacked under every Chinese character, and a translation into the language you pick. Tap any word for its pinyin and meaning. Works on YouTube, Netflix, and any site with real subtitle tracks — the extension detects them automatically. On YouTube, turn on Karaoke and the word being spoken is highlighted as you hear it.
2) Free — subtitles printed into the picture. No key, no cost. Chinese dramas, vlogs and Shorts often have no subtitle file at all — the words are burned into the image. The extension reads them with OCR that runs on your own computer, then adds the same Pinyin, translation and dictionary. No video frame ever leaves your device. It takes ~15–30 seconds to start, then runs live.
3) Anything with sound — built from the audio. Including videos with no subtitles at all, on any website. It listens to the tab's audio and creates the captions itself, auto-detecting the spoken language and translating live — so it works even where no subtitles exist. That includes live meetings: open Zoom, Google Meet or Microsoft Teams in a browser tab and it captions and translates the participants you hear, in real time — no bot joins your call. This mode is optional and runs on your own Soniox key.
If the words are on screen at all, it's free. They come from the site's own subtitle track (YouTube, Netflix, or any site with standard subtitles) — or, when the subtitles are printed into the picture, they are read off the video image by OCR running on your own computer. Either way the Pinyin and the dictionary also run entirely on your device, and no video frame is ever uploaded — no key to create. The translated line is free too: on YouTube it uses YouTube's own translation; on Netflix and other sites the subtitle text (never any audio or images) is sent to Google Translate for the translated line only. See the privacy policy for exactly what goes where.
When there are no words on screen to read — a video with no subtitles, a live meeting (Zoom, Meet or Teams open in a browser tab), or any other website — the audio route takes over. That's powered by the Soniox API with your own key, so you pay Soniox directly: pay-as-you-go, currently about US$0.12 per hour of audio (new accounts may get free starter credits). Creating a key takes a minute, with a link and steps in the extension's guide. To save credits it pauses the connection during silence and reconnects when sound returns.
The extension runs no server and has no analytics. Your key, if you add one, is stored only in your own browser.
Feedback is very welcome — it genuinely shapes what gets built. Reach out at hexaoutsourcing@gmail.com (click to copy).