Inspiration
Learning a new language is exciting, but that excitement doesn't always last. Many people start learning a language out of curiosity, only to abandon it a few weeks later when lessons become repetitive and motivation fades.
I wanted to explore a different approach. What if learning a language felt less like completing homework and more like catching up with a friend?
That's where Mi Solecito came from: a little sunshine that makes learning a language something you actually look forward to.
What it does
Mi Solecito is an AI-powered language learning companion that turns everyday moments into opportunities to learn.
Instead of relying entirely on traditional vocabulary exercises, users can interact with an AI companion through a natural calling experience, practice conversations, and learn practical words based on the world around them.
The goal is simple: make language learning more personal, interactive, and enjoyable, without the pressure of daily streaks or commitments.
How It is built
I designed Mi Solecito around a real-time conversational experience.
- LiveKit: Real-time audio/video communication infrastructure.
- Google Gemini: AI capabilities for understanding and responding to users.
- Frontend: The interface for interacting with the language-learning companion.
The architecture brings together real-time communication and conversational AI to create a more natural learning experience.
Challenges I ran into
One of the biggest challenges was figuring out how to make language learning feel like a natural conversation rather than another chatbot interaction.
Integrating real-time communication with AI also introduced technical challenges around connecting media streams, handling credentials, and coordinating the different services.
Another challenge was balancing simplicity with usefulness. I wanted the experience to be easy to start using while still providing meaningful learning opportunities.
Accomplishments
- Turning a familiar activity, making a phone call, into a new way to practice languages.
- Exploring a more conversational alternative to traditional vocabulary exercises.
- Designing an experience that prioritizes consistency and enjoyment over streaks and pressure.
- Bringing together real-time communication and AI to rethink how people interact with language-learning tools.
What I learned
I learned that building an AI-powered application involves much more than connecting an LLM to a frontend.
Real-time communication introduces a different set of engineering challenges, particularly when coordinating audio, video, and AI interactions.
More importantly, I learned to think beyond what AI can do and focus on how it can solve an actual user problem.
What's next for Mi Solecito
- Make conversations more personalized based on each learner's proficiency.
- Expand visual learning so users can discover vocabulary from objects in their surroundings.
- Introduce contextual exercises that reinforce words users encounter in everyday life.
- Improve conversational memory so Mi Solecito can build on previous interactions.
- Support more languages and make the experience accessible to more learners.
Ultimately, I want Mi Solecito to become the language-learning companion people actually look forward to calling.
Built With
- chert
- computervision
- facetime
- gemini
- livekit
- node.js
- realtimeai
- typescript
- voiceai
- webrtc
