VakyaAI is a multilingual language learning platform integrated with an AI-powered chatbot designed to enhance user engagement, accessibility, and personalized learning in education.
Leveraging state-of-the-art Natural Language Processing (NLP), LangGraph-based conversational memory, and low-latency inference with the Groq-hosted LLaMA 3.2B model, VakyaAI facilitates real-time, interactive conversations with grammar correction, vocabulary assistance, and contextual feedback.
A unique feature of VakyaAI is its seamless support for both text and voice-based interactions, enabling learners to practice speaking, listening, reading, and writing skills in multiple languages.
- Multilingual AI Chatbot — Supports multiple languages with contextual understanding.
- Voice & Text Interaction — Speak or type; receive both audio and text responses.
- Grammar & Vocabulary Assistance — Instant corrections and suggestions.
- Conversational Memory — Context-aware, coherent dialogues powered by LangGraph.
- Lightweight Deployment — Built using Streamlit for web-based interaction.
- Low Latency AI Inference — Powered by Groq-hosted LLaMA 3.2B for real-time responses.
- Self-paced Learning — Adaptive, interactive, and user-centric educational environment.
- Language learning for students, travelers, and professionals.
- Pronunciation practice through speech-to-text and audio feedback.
- Conversational AI for educational institutions.
- AI-assisted tutoring in multilingual settings.
| Component | Technology/Library | Purpose |
|---|---|---|
| Frontend/UI | Streamlit | Simple web app interface |
| AI Model | Groq (Llama 3 70B via Groq API) | Generates AI chat responses |
| Text-to-Speech (TTS) | gTTS (Google Text-to-Speech) | Converts text responses to speech |
| Speech-to-Text (STT) | SpeechRecognition + Google API | Converts voice input to text (local only) |
| Environment Vars | python-dotenv | Loads API keys from .env securely |
| Audio Encoding | base64 | Embeds audio players in the web page |
| Utilities | uuid, datetime, re | Manage session IDs, time stamps, text cleanup |
| Deployment | Streamlit Community Cloud / GitHub | Host app online |
This is how it works when you use VakyaAI as a mobile app.
Click the image above to watch the demo video showcasing the multilingual language learning mobile app in action.
This is how it works when this repository is cloned into your local system and run.
Click the image above to watch the demo video showcasing the AI-powered voice chatbot running locally.
- Clone the repository:
git clone https://github.com/your-username/VakyaAI.git
cd VakyaAI
pip install -r requirements.txt
streamlit run app/main.pyHow to Use VakyaAI Open the web interface. Select your preferred language. Choose Text Chat or Voice Chat. Start interacting with the AI — get grammar corrections, vocabulary suggestions, and pronunciation feedback.
contributions: Nandeesh C M [email protected]
Contact 📧 [email protected] 📧 [email protected]