🤖 Domain-Restricted Q&A Assistant

Domain-Restricted Q&A Assistant is a Streamlit-based AI application that answers questions only within a selected domain using a custom CSV knowledge base. The app combines OpenAI’s GPT models with domain restrictions and personalized response styles to create a focused AI assistant experience.

---

🛠️ Technologies Used

• Python
• Streamlit
• Pandas
• OpenAI API
• CSV-based Knowledge Base

---

🚀 Features

• Domain-specific AI assistant

• CSV knowledge base upload system

• AI responses powered by OpenAI

• Domain restriction enforcement

• Adjustable response styles:

  • Tone
  • Length
  • Audience level

• Quick prebuilt questions

• Session state persistence

• Error handling for invalid uploads and API issues

• Simple and responsive Streamlit UI

---

⚙️ Process

User selects a domain from the Setup tab

User uploads a CSV knowledge base containing:

• topic
• information

The application:

• Loads and validates the CSV
• Converts the CSV into a formatted knowledge base
• Stores it in Streamlit session state

User enters:

• A question
• Tone preference
• Response length
• Audience level

The app:

• Builds a structured prompt
• Sends it to OpenAI GPT-4o-mini
• Displays the AI-generated answer

Users can also select quick template questions from the Quick Questions tab.

---

🧑‍💻 How I Built It

I built this project using Streamlit as the frontend framework and OpenAI’s API for AI-generated responses.

The development process involved:

• Designing a multi-tab Streamlit interface
• Creating domain-restricted prompt engineering
• Building CSV upload and validation functionality
• Managing persistent state using Streamlit session state
• Integrating OpenAI API calls
• Adding customizable response styles
• Creating reusable helper functions
• Implementing quick-question templates
• Handling file and API errors gracefully

---

📚 What I Learned

• How Streamlit session state works
• Working with CSV uploads in Streamlit
• Prompt engineering for domain restriction
• Integrating OpenAI APIs into Python applications
• Structuring reusable functions and clean app architecture
• Error handling for file uploads and API responses
• Building multi-tab interactive applications
• Managing dynamic UI components in Streamlit

---

📈 Overall Growth

This project improved my understanding of AI application development, prompt engineering, and frontend-backend interaction using Streamlit.

It also helped me practice:

• Clean function-based architecture
• API integration
• State management
• User-focused UI design
• Building domain-constrained AI systems

---

🔧 What Could Be Improved

• Add conversation history
• Support multiple uploaded knowledge bases
• Add embeddings/vector search for smarter retrieval
• Improve prompt engineering
• Add authentication system
• Export chat responses
• Add dark mode support
• Improve UI styling and animations
• Add real-time streaming responses

---

▶️ How to Run It

git clone <your-repo-url>
cd domain-qa-assistant
pip install -r requirements.txt
streamlit run app.py

Then open:

http://localhost:8501

---

📂 Required CSV Format

Example CSV structure:

topic,information
Protein,Protein helps build muscle.
Cardio,Cardio improves heart health.
Budgeting,Budgeting helps manage expenses.

Required columns:

• topic
• information

---

🔑 API Setup

Enter your OpenAI API key in the sidebar after launching the application.

Get your API key from:

OpenAI Platform