Mux | Devpost

Inspiration

Anyone who's tried learning a new language knows that it's a difficult process, especially learning how to speak it because there's such limited opportunity for live interaction and feedback. There's no shortage of resources for reading or writing, but actually having conversations? That's where it gets challenging. Each of us on the team has faced that frustration. We've all tried picking up different languages and wishing that you had someone to practice with, someone who could guide you and offer real feedback without the pressure. What if that solution is one call away?

What it does

Mux is an AI language partner, who's available at the press of a button. Anyone with a phone can access it at any time. Users call in and engage in real-time dialogue. Mux can offer real time news, play games to deepen understanding, and practice live speaking scenarios to expand your vocabulary. It listens, responds, and adapts—offering feedback that helps you improve with every interaction. Mux turns language learning into a human experience, accessible anytime, anywhere.

How we built it

We built Mux using Python, and leveraged Twilio as the interface to bridge OpenAI's advanced language model with live phone interactions. This integration allowed us to handle voice recognition and response, transforming a simple phone call into a dynamic learning experience. We also integrated a news feature that used the NewsAPI to provide news stories, so users can practice their language skills while staying informed about current events. OpenAI’s voice features enabled natural conversation and voice recognition for a truly interactive experience. Each component worked together to make Mux an assistant that speaks, listens, and learns, all in real time.

Challenges we ran into

Building Mux meant pushing the limits of real-time interaction. Integrating voice feedback and natural conversation while maintaining high performance was a challenge. We had to innovate around processing speed, accuracy, and memory. Teaching Mux to understand and adapt to each user's learning style added another layer of complexity, as we wanted Mux’s feedback to be not just accurate but also insightful and encouraging. Also constantly, running out of credits, and reaching the user limits on APIs.

Accomplishments that we're proud of

We’re proud of creating something that doesn’t just respond but actually engages, challenges, and teaches. Most importantly, we built something that makes language learning enjoyable. Mux feels alive. When you talk to it, you’re not just hearing responses—you’re experiencing a real connection with language.

What we learned

Creating Mux taught us a lot about the complexities of natural language processing and the intricacies of user interaction. We learned how critical feedback and real-time engagement are to effective learning. And, on a deeper level, we learned the importance of designing an experience that feels personal—because learning is ultimately about connection.

What's next for Mux

We’re planning to expand its capabilities by incorporating more cultural content and deeper contextual feedback. We want Mux to be personalized for each user, and after every conversation. In the future, Mux will be able to detect learning patterns and dynamically adjust to support each learner’s unique journey to fluency.