Repolode
Submit
niyongsheng

niyongsheng/AudioScript

A macOS desktop app that captures system audio and shows live bilingual English-Chinese subtitles, with speech recognition and translation running fully on-device.

Open-source alternative to:Xunfei Simultaneous Interpretation

View on GitHubWebsite
Stars
47
Forks
9
Open issues
0
Last commit
2 days ago

What is it?

AudioScript listens to system audio output and turns foreign-language podcasts and videos into live subtitles, showing the original text alongside a Chinese translation in a floating window. It is aimed at users who follow international meetings, English podcasts, or untranslated video without relying on a cloud service.

Speech recognition uses sherpa-onnx in streaming mode, and translation uses CTranslate2 with opus-mt models for English and Chinese in both directions. Audio can come from system sound, the microphone, or both mixed. The subtitle window shows bilingual lines, and font size, opacity, and line count are adjustable.

The app is built with Tauri 2 and targets macOS 14.2 or later. It runs in the menu bar tray, keeps running when the window is closed, allows a single instance, and can start at login.

Key features

  • 01Captures system audio, microphone, or both at once.
  • 02Streams speech recognition with sherpa-onnx.
  • 03Translates English and Chinese offline with CTranslate2 and opus-mt.
  • 04Shows bilingual subtitles in an adjustable floating window.
  • 05Runs in the menu bar tray with optional launch at login.

Use cases

  • Following an English conference or webinar live in Chinese.
  • Understanding English podcasts while commuting or working.
  • Watching foreign-language videos without official subtitles.
  • Keeping audio processing local when cloud transcription is not acceptable.

Quick start

  1. 1

    Download the app from the releases page

  2. 2

    Install dependencies and the ASR model

    pnpm install
  3. 3

    Fetch the translation model

    pnpm fetch:mt
  4. 4

    Start in development mode

    pnpm tauri dev

Things to know

  • Only runs on macOS 14.2 or later.
  • Translation covers English and Chinese only (en ↔ zh).

Topics

#speech-recognition#real-time-transcription#translation#subtitles#offline#tauri#macos#sherpa-onnx

FAQ

What is AudioScript?
A macOS desktop app that captures system audio and shows live bilingual English-Chinese subtitles, with speech recognition and translation running fully on-device. AudioScript listens to system audio output and turns foreign-language podcasts and videos into live subtitles, showing the original text alongside a Chinese translation in a floating window. It is aimed at users who follow international meetings, English podcasts, or untranslated video without relying on a cloud service.
Is AudioScript free to use? What license does it have?
AudioScript is open source under the MIT license. Check the license terms before using it in commercial products.
Is AudioScript still maintained?
Actively maintained. The last commit to AudioScript was on Oct 19, 2026.
What can I use AudioScript for?
Following an English conference or webinar live in Chinese. Understanding English podcasts while commuting or working. Watching foreign-language videos without official subtitles. Keeping audio processing local when cloud transcription is not acceptable.
How do I install AudioScript?
Download the app from the releases page Install dependencies and the ASR model: pnpm install Fetch the translation model: pnpm fetch:mt Start in development mode: pnpm tauri dev
Where can I get AudioScript?
The source code and installation instructions are on GitHub: https://github.com/niyongsheng/AudioScript

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects