---

PsyAssist: Multimodal RAG-Based Mental Health Chatbot

It is an intelligent, context-aware mental health chatbot that leverages Retrieval-Augmented Generation (RAG) built on a data-driven vector database. PsyAssist uses a custom mental health knowledge base in PDF format to provide personalized, empathetic support through tonality-aware dialogue. Built using Flask, it features a smooth and accessible web interface.

---

Features

1. Retrieval-Augmented Generation (RAG)

Uses a vector database created from mental health documents to retrieve contextually relevant answers tailored to user questions in a data-driven manner.

2. Tonality-Aware Conversations

Dynamically adjusts responses based on the emotional tone detected in user input, enhancing empathy and relevance.

3. Multimodal Framework

Engineered to support diverse inputs (text and easily extendable to voice/image) for a richer conversational experience.

4. Clean Web Interface

Built using HTML/CSS/JS to provide a seamless and user-friendly environment focused on wellness and support.

---

Interface Snapshots

Chat Interface	Help/FAQ Page

---

Project Structure

├── index.html          # Main chatbot interface
├── help_fnq.html       # Help / FAQ page
├── style.css           # Frontend styling
├── script.js           # Client-side interactivity
├── app.py              # Flask backend logic
├── rag_embeddings.py   # PDF embedding and vector store
├── resources/          # Static assets
└── rag_database/       # (Create this) Add mental health PDFs here

---

Getting Started

1. Clone the Repository

git clone https://github.com/your-username/PsyAssist.git
cd PsyAssist
mkdir rag_database
pip install -r requirements.txt

2. Configure API Keys

Add any necessary API keys (e.g., for embedding models) in your environment or directly in the embedding script.

3. Prepare Knowledge Base

Place relevant mental health PDF documents in the rag_database/ directory. These documents will be processed into vector embeddings for retrieval.

4. Launch the Application

python app.py

Visit http://localhost:5000 in your browser to start chatting.

---

Contributing

You're welcome to contribute by improving features, UI, or expanding the knowledge base:

• Fork the repository
• Create a feature branch
• Submit a pull request with a clear explanation

---

Author

Soumya Sourav Das

Portfolio | GitHub | LinkedIn

---