The app basically works this way:

* You listen to podcasts shows and YouTube channels of content you really care about, content you would watch normally in your native language.

* After each sentence, you try to imitate the native speaker, using the language shadowing technique. [1]

* You click a word to learn its meaning and usage in context, and mark words as known as you get exposed to more content.

The main idea is to immerse yourself in content you care about, and learn with context. These are some of the techniques I’ve used before and I think are the best “bang for your buck” in terms of time invested. This is based on my experience, of course there can be better ways to become fluent faster, let me know.

Originally the app also included an “AI tutor”, a way to write and have a conversation in your target language. I found that I didn’t use the feature myself that much so I removed it. Let me know if you find value in such a feature and I can bring it back.

I tried to avoid “gamification”, but I included some stats that can help you track your progress.

Some technical challenges I found while building this:

* Most podcasts today use “dynamic ad insertion”, meaning that a podcast can have different content for the same URL, depending on the IP of the request or the time, that means I have to store a copy of the podcast so that the transcription matches to all the users.

* Originally I used Stripe, but Stripe is not a “merchant of record” (meaning they don’t collect taxes on your behalf) so I had to switch to Lemon Squeezy.

* I started building this using Nextjs 13 when the app router was not production ready, I ran into some issues with their aggressive caching strategy and decided instead to redo the whole thing and launch a Web + IOS + Android app using react native, sharing most of the codebase. I ran into issues there related to SSE on react native, and decided to just go back to Nextjs (now stable) and just launch it. (Moral of the story … just launch it :) )

So let me know what you think. I have been using it myself so far, so any real user feedback is valuable.

[1] https://en.wikipedia.org/wiki/Speech_shadowing#Language_lear...