💸 AI-Powered Financial Assistant

An intelligent AI Financial Assistant built for the Google Hackathon!
This tool helps users simplify money management by offering real-time insights, PDF-based financial summaries, and budgeting suggestions — all through natural conversation.

---

🔧 Features

• ✅ Conversational AI with Gemini for smart financial dialogue
• 🔐 Firebase Authentication for secure and easy user login
• 📄 Retrieval-Augmented Generation (RAG) with external APIs
• 🌐 WebSocket-based Real-Time Communication for dynamic, streaming chat
• 💬 Modern Chat UI with React
• ⚡ FastAPI Backend for performance and simplicity

---

🔍 Tech Stack

Layer	Technology Used
💬 Frontend	React, Tailwind CSS
⚙️ Backend	FastAPI, Python, WebSockets
🔒 Auth	Firebase Authentication
🧠 AI Engine	Gemini Pro (Google AI Model)
📄 RAG	Sklearn

---

🔐 Firebase Authentication

Firebase handles secure authentication with:

• ✅ Email/Password Login
• 🔗 Google Sign-In
• 🐙 GitHub Login

---

🧠 GenAI Financial Assistant (Backend)

This backend powers a conversational AI-driven financial advisor bot that provides tailored investment advice using Google Gemini Pro, Firebase, WebSockets, and Retrieval-Augmented Generation (RAG) for enhanced knowledge grounding.

---

🚀 Features

🔐 Firebase Authentication

• Supports Email/Password, Google, and GitHub login.
• Firebase Admin SDK handles secure user creation and token verification.
• Firestore is used to store user profiles and preferences.

🧠 Conversational AI with Gemini

• Uses Google's Gemini 1.5 Pro model (google.generativeai) for natural, personalized financial advice.
• Adapts advice based on:
  • User's investment experience level (Beginner, Intermediate, Expert)
  • Preferred investment type (e.g., Long-term, Passive income, etc.)
• Responses include:
  • TL;DR summaries
  • Risk assessments
  • Pro tips & common pitfalls
  • Step-by-step action plans
  • Emoji-enhanced answers for better readability
• System prompt includes behavioral instructions for humor, analogies, and follow-up questioning

📚 RAG (Retrieval-Augmented Generation)

• Relevant documents are retrieved using a custom rag_module to ground Gemini responses in real facts.
• Provides better accuracy and reliability for advice and definitions.

🌐 WebSockets

• Real-time profile change monitoring using WebSocket endpoints.
• Backend pushes updates to connected clients whenever Firestore user data changes.

🧩 Session Handling

• Short user prompts are automatically merged with previous ones for context retention.
• Each user has an active session maintained in memory to simulate multi-turn conversations.

---

📦 API Endpoints

GET /

Home route – "Welcome to GenAI Financial Assistant!"

GET /healthcheck

Returns service health status.

POST /signup

Creates a new user in Firebase and stores profile in Firestore.

POST /login

Frontend handles login using Firebase SDK; this route is informational.

POST /update-profile

Updates user's investment type and experience level.

POST /recommend

Core route for generating investment advice using Gemini + RAG + profile data.

GET /ws/{user_id}

WebSocket endpoint to push real-time updates to specific users.

---

🛡️ Security

• Auth middleware verifies Firebase tokens using HTTPBearer scheme.
• All protected routes require a valid JWT from Firebase.

---

⚙️ Tech Stack

Tech	Usage
FastAPI	Backend framework
Firebase	Auth, Firestore, Realtime updates
Gemini Pro	Financial advice generation (GenAI)
RAG	Knowledge grounding via custom module
WebSockets	Real-time client updates
CORS	Enabled for localhost:3000 (frontend)

---

frontend

💸 GenAI Financial Assistant – Frontend

This is the frontend for the GenAI-based Financial Assistant. It provides a sleek chat interface for users to interact with the assistant, which offers financial recommendations using GenAI and a FastAPI backend.

---

🔧 Tech Stack

• Next.js (React Framework)
• TypeScript
• Firebase Authentication
• Tailwind CSS
• React Icons
• React Markdown
• FastAPI (Backend Integration)

---

🚀 Features

• Google and GitHub authentication using Firebase
• Chat interface with real-time user-bot conversation
• Markdown rendering for bot responses
• Multiple chat sessions with chat history
• Integration with backend /recommend endpoint for financial recommendations

---

🛠️ Setup & Run Locally

1. Clone the repository

bash git clone https://github.com/your-username/GenAI-Financial-Assistant.git cd GenAI-Financial-Assistant/frontend

2. Install dependencies bash npm install

3. Create .env.local file bash touch .env.local

4. Add your Firebase configuration:

env NEXT_PUBLIC_FIREBASE_API_KEY=your_api_key

NEXT_PUBLIC_FIREBASE_AUTH_DOMAIN=your_auth_domain

NEXT_PUBLIC_FIREBASE_PROJECT_ID=your_project_id

NEXT_PUBLIC_FIREBASE_STORAGE_BUCKET=your_storage_bucket

NEXT_PUBLIC_FIREBASE_MESSAGING_SENDER_ID=your_sender_id

NEXT_PUBLIC_FIREBASE_APP_ID=your_app_id

5. Run the development server

npm run dev

Open your browser at http://localhost:3000

🔗 Backend Integration

Ensure your FastAPI backend is running locally at:

bash http://localhost:8000/recommend

The frontend sends authenticated POST requests to this endpoint using Firebase ID tokens.

🛠️ Scripts

Script	Usage
npm run dev	Start the Next.js app in development mode at localhost:3000
npm run build	Create an optimized production build
npm run start	Start the production server
npm run lint	Run ESLint to check code formatting and quality
npm install	Install project dependencies from package.json

Check out our Next.js deployment documentation for more details.