KaitaКайта · 凯塔

Hear the phrase.
Say it again.

“Kaita” is Kyrgyz for “again”.

Kaita turns any YouTube video with captions into a Chinese speaking and interpreting gym. A practice panel docks next to the video: scored shadowing, phrase-by-phrase and simultaneous interpreting.

Add from Chrome Web Store

Coming soon: Kaita is under review in the Chrome Web Store.

For Google Chrome and Microsoft Edge on desktop. Free, no sign-up.

See how it works

Phrase 3 of 80:08

这zhè是shì第dì二èr十shí五wǔ课kè。

This is lesson twenty-five.

You said

这是第二十五课。

86 of 100

Nice: 6 of 7 characters matched. 课 kè is fourth tone — keep it short and falling.

Four modes in one panel

You train on real native speech — lectures, news, interviews, dramas — instead of textbook dialogues. Switch modes with the same tabs the panel uses.

Watch: captions that follow the video

Captions scroll in sync with the video and the current phrase is highlighted. Star any line to come back to it later.

Shadowing (跟读)

The phrase plays, the video pauses and the mic opens automatically. Repeat after the speaker and Kaita shows which characters matched, gives a 0–100 score and a gentle pronunciation tip: tones, zh/z, an/ang and other common slips.

The Kaita panel next to a video: the phrase 这是第二十五课。, a mic button and a score of 100.

Consecutive interpreting

Hear a phrase, say your translation aloud, compare it with the reference and rate how hard it was.

Consecutive mode: your spoken translation shown next to the reference translation.

Simultaneous interpreting (同声传译)

The video keeps playing while you interpret. Afterwards you get an “original | your version | reference” table and a coverage score.

Simultaneous session summary: 67% coverage and a table of original, your version and reference.

Library and review

Saved phrases, video history, attempts and interpreting sessions live on one Library page: search, spaced review, daily stats, CSV and JSON export. “AI analysis” builds a ready-made prompt from your mistakes that you can paste into any AI chat. Kaita itself sends nothing anywhere.

The Library page: saved phrases with scores, difficulty ratings and export.

Who it’s for

  • Mandarin learners

    From HSK 2–3 to advanced: listening, tones and fluency.

  • Interpreting students and interpreters

    A daily consecutive and simultaneous warm-up on authentic material.

  • Chinese speakers learning Russian

    It works both ways, Chinese ↔ Russian, and English is supported too.

Get started

  1. Install Kaita

    From the Chrome Web Store, in Chrome or Microsoft Edge.

  2. Open a YouTube video

    One with Chinese or Russian captions.

  3. Click the Kaita icon or press Alt+E

    The panel opens next to the video.

  4. Allow the microphone

    On the first recording the browser asks for mic access for www.youtube.com: recording happens on the video page.

Add from Chrome Web Store

Coming soon: Kaita is under review in the Chrome Web Store.

Your voice stays on your computer

On-device recognition
The built-in engine is the open Whisper model running in your browser. It downloads once (about 80 MB); after that recognition works offline.
No server of ours
No sign-up, ads or analytics. Phrases, scores and settings stay in your browser.
Your choice in Chrome
Edge always uses Whisper. In Chrome, “Auto” uses Chrome’s own speech recognition (a Google service); pick “Кайта · Whisper” in settings to keep your voice on your device.

Privacy policy

Pricing

Available now

Beta

$0 now

  • All four modes and the Library
  • On-device Whisper speech recognition
  • No sign-up, no card

Subscription

$5 per month, after beta

  • Free trial
  • Price and features shown at checkout before you pay
  • We’ll announce it in the extension in advance

Questions

Is it free?

Yes — every feature is free during the beta. Later there will be a $5/month subscription with a free trial.

Which videos work?

Any YouTube video with captions — Chinese, Russian or English. Without captions Kaita has nothing to compare your speech with.

Where does my voice go?

The built-in “Кайта · Whisper” engine recognises speech on your computer, and Kaita has no server. The exception is Chrome’s “Auto” or “Browser” setting, where Google’s service does the recognition. See the privacy policy.

Does it work in Microsoft Edge?

Yes. Open Kaita’s Chrome Web Store page in Edge and allow extensions from other stores — here’s how.

Can I use it on my phone?

No. Extensions run in desktop Chrome and Edge (Windows, macOS, Linux).

What language is the interface in?

Russian for now. More interface languages are in progress.

How do I keep my progress?

All data lives in your browser. Before removing the extension or moving to another computer, export to JSON on the Library page and import it later.