Hey everyone,

Like many here, I rely heavily on reading and listening for immersion, but I’ve always been frustrated by the existing options. LingQ has gotten absurdly expensive ($15/month) for software that feels slow and cluttered, and most free alternatives lack a seamless reading/listening workflow.

So I decided to build Lumen, a native, lightweight, and completely free desktop app designed specifically for immersion learning.

What it does:

• Zero latency reading: Built with Tauri 2 and WebKit (<100MB footprint, no Electron bloat). Clean serif typography and minimal UI so you actually focus on reading.
• Instant lookup & SRS grading: Click any word or highlight a phrase to see contextual meanings and save them to your vocabulary list.
• Integrated YouTube & article importer: Search and watch YouTube videos directly inside the app, convert them into lessons, and read along with synchronized audio.
• 100% local-first audio & AI:
  • Ultra-fast offline TTS using Supertonic 3 (sub-millisecond natural audio for words/phrases).
  • Local ASR via Whisper / Qwen3-ASR for perfect audio/text alignment (with an audio resync button if drift occurs).
  • Offline grammar/translation support via quantized local LLMs (Qwen 3.5), or the option to plug in your own DeepSeek API key for dirt-cheap, highly nuanced context translation.
• Anki-friendly: Easy CSV export with your target word, meaning, and the exact sentence context.

Why free / open source?

Language tools should be accessible. The app runs locally on your machine, stores everything in a local SQLite database, and respects your privacy.

The code is open source, and pre-built binaries (macOS / Windows) are available on GitHub:

https://github.com/LivingTwice/lumen-releases

It’s still an active project that I use daily for my own Italian immersion, so I’d love to get your feedback, bug reports, or feature requests. Hope it helps some of you streamline your input routine!