SAHAAYI

Enhancing Personalized Learning Experiences

SAHAAYI is an innovative platform designed to provide personalized learning experiences for differently-abled individuals. By integrating advanced AI/ML models and accessibility-focused features, it ensures inclusivity and adaptability for all users.

Key Features

1. Braille Conversion

• Converts text to Braille using Unicode mapping.
• Fetches video transcripts from YouTube using YouTubeTranscriptAPI.
• Outputs original and Braille transcripts.

Workflow:

1. Accepts YouTube video URL.
2. Extracts video ID and retrieves transcript.
3. Converts transcript to Braille for accessibility.

---

2. Hand Gesture Detection and Recognition

• Real-time hand tracking and gesture recognition using Mediapipe.
• Supports accurate detection of finger positions for sign classification.
• Aids communication through hand-sign tutorials.

Dependencies:

• OpenCV for image processing.
• Mediapipe for hand tracking.

---

3. MCQ Generation

• Generates high-quality, topic-specific MCQs using Gemini AI.
• Provides real-time feedback and score tracking.
• Features hover-to-speak functionality for enhanced accessibility.

Workflow:

1. User inputs a topic.
2. System generates structured MCQs.
3. Displays questions in an interactive quiz interface.

---

4. Retrieval-Augmented Generation (RAG)

• Uses semantic search to retrieve relevant knowledge.
• Synthesizes coherent responses with Google Gemini AI.
• Combines dense embeddings (SentenceTransformer) and FAISS for efficient retrieval.

---

Tech Stack

Frontend

• React with NextJS
• TailwindCSS for design
• ShadCN components

Backend

• Python, Django, and Django Rest Framework

AI/ML

• Transformers model (all-MiniLM-L6-v2)
• FAISS for similarity search
• Gemini LLM 1.5 Flash

---

Installation

1. Clone the repository:

   git clone https://github.com/your-repo/SAHAAYI.git

2. Install dependencies:

   npm install  # For frontend  
   pip install -r requirements.txt  # For backend

3. Set up the environment variables as required.

4. Start the development servers:

   npm run dev  # Frontend  
   python manage.py runserver  # Backend