I’m the technical half of a two-person team. I partnered with an English teacher (1.5M followers) to build a voice-first language learning app.

The goal was to solve the "Intermediate Plateau". Most apps (like Duolingo) are great for vocabulary, but they gamify the process so much that users become addicted to "streaks" rather than actually speaking.

*The Experiment (Dev + Creator):* We have $0 marketing budget. I handle the product/engineering, he handles the distribution. We are betting that "Product + Audience" is enough to compete with VC-backed giants.

Stack:

- Mobile: React Native (Expo). We needed to ship iOS and Android fast with a single codebase. - Voice: 11Labs. We cloned the partner's voice. The main engineering challenge is minimizing latency to make the conversation feel natural (we are currently optimizing the stream buffer). - Logic: OpenAI (Custom system prompts for roleplay scenarios like "Job Interview" or "Customs Officer"). - Backend: Supabase.

We launched few days ago after getting rejected 10 times in 14 days. We were stuck in a loop with Design Guidelines for the Paywall. Imo they were not checking the updates as I was not submitting a new build (it was handled by RevenueCat), I had to textually explain that for them to approve the app.

I’m happy to answer questions about the React Native <-> 11Labs integration, the latency challenges, or the partnership structure.

PS: If you want to try it, this link gives 50% off the first month: https://apps.apple.com/redeem?ctx=offercodes&id=6755333340&c...