The AI Tutor for Underrepresented Languages
"Your new go-to language learning web platform for underrepresented languages."
Team: Aymen Saya, Zaynah Wahab, Tisha Kaur, Josslyn Quansah
Vernacular is an interactive web-based language learning platform that leverages AI to preserve and teach underrepresented languages.
Our first implementation focuses on Bengali, integrating conversation, pronunciation, and vocabulary tools into a single immersive experience.
The project uses AI-powered tutoring, speech recognition, and text-to-speech to create an engaging, accessible learning journey with the long-term goal of expanding to other languages like Punjabi, Urdu, Twi, and Swahili.
- Explore essential vocabulary and phrases in Bengali.
- Listen to correct pronunciations using Text-to-Speech.
- Practice with built-in recording and feedback.
- Chat with an AI Bengali Tutor to improve fluency.
- Engage in natural text or voice conversations.
- Ask about grammar, meaning, or pronunciation.
- Automatically add new vocabulary to your custom word bank.
- Save words you’ve learned locally.
- Interactive feedback with playful, colorful UI.
| Layer | Technologies |
|---|---|
| Frontend | HTML, CSS, JavaScript |
| Backend | Python, Flask, Flask-CORS |
| AI/ML | Google Gemini API for natural language understanding |
| Speech/Audio | Web Speech API (Text-To-Speech + Speech Recognition) |