jasoncheng7115

jasoncheng7115 / jt-live-whisper

Public

100% 全地端 AI 語音工具集：即時轉錄、即時翻譯、錄音檔批次處理、講者辨識、會議摘要，所有 AI 模型皆在自有設備上運行，資料不經過任何雲端服務。

102
18
100% credibility
Found Mar 09, 2026 at 102 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A macOS tool for capturing system audio to provide real-time speech transcription, English-Chinese translation, speaker diarization, and AI-generated meeting summaries using only local AI models.

How It Works

1
📰 Discover the tool

You hear about this Mac app that instantly translates English meetings or videos into Chinese subtitles, all without sending your words anywhere.

2
📥 Easy one-click install

Copy a simple line into your terminal, and it downloads everything needed to get started in minutes.

3
🔊 Set up sound sharing

In your Mac's sound settings, create a mixer so the app can hear audio from Zoom, YouTube, or any video playing.

4
🚀 Launch and choose mode

Run the starter, pick live translation or file processing, and watch it come alive with your choices.

5
🎤 See subtitles appear

As voices play, real-time Chinese text scrolls with timings, speaker colors, and smooth translations.

6
Handle saved audio
✨
Get smart summary

AI pulls key points and timelines into neat, colored notes.

✅
Save and review

Export transcripts with playback to relive meetings easily.

🎉 Private, perfect notes

Enjoy clear understanding of any talk, with everything staying safe on your Mac—no costs or clouds.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 102 to 102 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is jt-live-whisper?

This Python toolkit delivers 100% local AI speech processing on macOS: live transcription and translation (English-Chinese or vice versa), batch audio file handling, speaker diarization, and meeting summaries—all powered by offline Whisper models. It captures system audio from any source like Zoom calls, YouTube, or podcasts via a virtual device, spitting out real-time subtitles in your terminal or polished HTML transcripts with timelines. No cloud APIs, no data leaks, just your hardware doing the heavy lifting.

Why is it gaining traction?

Unlike cloud-dependent tools, it runs fully offline with zero API costs or privacy risks, handling everything from live en2zh subs (~300ms latency options) to diarized summaries via simple CLI flags like `--input meeting.mp3 --diarize --summarize`. One-click install scripts fetch models and deps, auto-detecting local LLMs like Ollama, and it scales to remote GPU servers for 5-10x faster batch jobs. Devs dig the app-agnostic audio grab and interactive menus that skip boilerplate setup.

Who should use this?

Non-native English speakers in tech meetings (e.g., devs on global teams juggling Zoom/Teams). Podcasters or researchers needing private, offline transcription of long recordings. macOS pros avoiding github 100 mb file limits on uploads by keeping everything local.

Verdict

Grab it if you need on-device speech AI now—docs are thorough, install is smooth, and core live Python Whisper flows work reliably. At 93 stars and 1.0% credibility, it's an early gem in the 100 github projects space: mature enough for daily use, but watch for edge-case polish as it grows.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.