✨ Inspiration
While using popular language learning apps, I noticed I was forgetting most of the words I had learned — simply because I wasn’t getting enough contextual practice.
Then I discovered something powerful: when I saw my known words in multiple sentences — and I already knew the surrounding words — learning became effortless and enjoyable.
That insight led me to start generating my own stories using only the words I had already learned.
It worked wonders for my memory and boosted my reading confidence.
For the past 8 months, I’ve been thinking about turning this into an app — and thanks to the Bolt.New Hackathon, I finally did it.
📚 What it does
Stolingo is a language learning web app that:
- Lets users learn vocabulary through interactive flashcards
- Tracks each word they've mastered
- Allows them to generate custom stories using only their learned words
- Provides a click-to-translate feature for instant word lookup
- Builds confidence through reading without unknown vocabulary
It’s like turning your vocabulary list into a personalized reading experience.
🛠️ How I built it
- Frontend & logic: Built with Bolt.New, allowing for a fast and visual development process
- Backend: Firestore (Firebase) was used for storing users, vocabulary, and story data
- LLM Integration:
- I used LLaMA 4 as the foundation model
- Served through Groq, the fastest LLM inference platform available today
- This enabled real-time story generation based on individual vocabulary data
- I used LLaMA 4 as the foundation model
- Story logic: The app dynamically crafts prompts using the user's current known word list and topic selection
🧱 Challenges I ran into
- Designing the story generation logic so that it strictly uses only learned words
- Structuring the Firestore database to manage word tracking and story linking effectively
- Balancing user-level customization with clean UX
- Crafting prompts that work reliably with LLaMA 4 while staying efficient in token usage
🏆 Accomplishments that I'm proud of
- Successfully integrated Groq + LLaMA 4 into a web app for personalized story generation
- Created a fully functional product that delivers meaningful value to language learners
- Deployed the app with a custom domain and real user experience — all during the hackathon
- Built everything solo, from concept to working product, in just a few days
📖 What I learned
- How to connect real-time LLM generation with user-specific educational data
- How to structure scalable Firestore schemas
- How incredibly powerful and fast Groq is for language learning use cases
- Bolt.New can turn solo ideas into live products blazingly fast**
🔮 What's next for Stolingo
- 🧩 Adding a fun, interactive module to make vocabulary learning more engaging and game-like
- 📱 Launching a mobile version for both iOS and Android to reach learners everywhere
- 🧠 Integrating the Mind Palace technique to boost long-term memory retention through visual association
- 💳 Adding a payment system to support a freemium model with unlimited story generation and premium features
- 🌍 Expanding multilingual support and allowing users to customize story tone and genre
- 🧠 Exploring smarter flashcard personalization based on story reading behavior
🦊 Stolingo is live on the web — and it’s just getting started.
Built With
- bolt
- entri
- firebase
- firestore
- groq
- javascript
- llama
- netlify
- react
- vite
Log in or sign up for Devpost to join the conversation.