We're Ryan and Dre, engineers who built Scena, a platform for creating and sharing interactive AI personas.
While learning Dutch and Spanish, we found language apps focused on gamification but lacked real conversation practice. So we built tools for users to create and share their own AI conversation partners.
The tech stack uses Azure speech services and fine-tuned GPT 4o mini, with a Swift frontend and AWS backend. What makes us different is treating personas like YouTube, anyone can create and share them. Early users have created everything from language learning through murder mysteries to technical interview practice.
Everything is free and open to try. We'd love feedback, especially on scaling user generated AI content.