#4747th of 49 in Desktop AI appsFor voice, speech and music

my-translator

Real-time speech translation overlay for system audio and mic on macOS and Windows

Stars
1.3k
License
MIT
Last commit
Jul 2026
Last release
Jul 2026
Language
Python
Stars, last 29 weeks+374
21 Apr 2026: 947 stars28 Apr 2026: 973 stars5 May 2026: 1,002 stars12 May 2026: 1,030 stars19 May 2026: 1,092 stars26 May 2026: 1,109 stars2 Jun 2026: 1,121 stars9 Jun 2026: 1,126 stars16 Jun 2026: 1,131 stars23 Jun 2026: 1,160 stars30 Jun 2026: 1,169 stars7 Jul 2026: 1,177 stars14 Jul 2026: 1,183 stars21 Jul 2026: 1,189 stars28 Jul 2026: 1,195 stars4 Aug 2026: 1,242 stars11 Aug 2026: 1,250 stars18 Aug 2026: 1,257 stars25 Aug 2026: 1,263 stars1 Sept 2026: 1,268 stars8 Sept 2026: 1,273 stars15 Sept 2026: 1,277 stars22 Sept 2026: 1,281 stars29 Sept 2026: 1,283 stars6 Oct 2026: 1,294 stars13 Oct 2026: 1,296 stars20 Oct 2026: 1,303 stars27 Oct 2026: 1,312 stars3 Nov 2026: 1,317 stars10 Nov 2026: 1,321 stars

947 on 21 Apr 20261,321 on 10 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Tauri 2 desktop app that captures system audio or microphone input, transcribes and translates it live, and shows the result in a minimal overlay. It supports four engines: Soniox, OpenAI Realtime, Qwen LiveTranslate Flash, and an offline MLX mode (Whisper + Gemma) on Apple Silicon. The app talks directly to the APIs you configure, with no relay server; optional TTS narration is available.

Who it is for: Users who need live translation of calls, talks or videos on their desktop

Strengths

  • Four engines, including a fully offline MLX mode on Apple Silicon
  • Two-way translation with speaker detection for Zoom, Meet and Teams calls
  • Direct API calls with local keys; no account, telemetry or relay server
  • Custom term glossary and three TTS providers (Edge, Google, ElevenLabs)

Weaknesses

  • Local mode is experimental, Apple Silicon only, about 10 s latency
  • Local mode covers only JA/EN/ZH/KO to VI/EN
  • OpenAI Realtime costs about $4/hr and disables two-way mode and custom TTS
  • Qwen engine needs explicit source language and Singapore-region key; no source transcript

What it needs

  • GPU optional
  • Needs Soniox API key, OpenAI API key, Alibaba DashScope API key, Edge TTS / Google Cloud TTS / ElevenLabs (optional)
  • Models: Soniox, OpenAI gpt-realtime-translate, qwen3-livetranslate-flash-realtime, MLX Whisper + Gemma

Quick answers

Is my-translator open source?

Yes. Its code is on GitHub under the MIT license.

Does my-translator need a GPU?

No. It runs on a CPU and can use a GPU when you have one.

Also in Desktop AI apps

See all 49
Also in Desktop AI apps
RankProjectScore
1cherry-studioDesktop client for chatting with many LLM providers, with agents and knowledge bases53.4k stars+1,124 (+2.1%) in 30 daysAGPL-3.092 out of 100
2dsh-desktopNative Windows and macOS desktop client for DeepSeek Harness, built on plugins35.7k stars+5,734 (+19%) in 30 daysMIT90 out of 100
3screenpipeRecords screen and audio locally, makes it searchable for you and your agents22.3k stars+481 (+2.2%) in 30 dayscustom license90 out of 100
#4prompt-optimizerPrompt rewriting and testing tool for web, desktop, Chrome and MCP Live demo ↗ (opens in a new tab)41.2k stars+3,703 (+9.9%) in 30 dayscustom license86 out of 100
#5aionuiDesktop and WebUI app that runs a built-in agent plus external CLI agents34.2k stars+923 (+2.8%) in 30 daysApache-2.086 out of 100