FinanceBrain AI

Real-Time Document Assistant for Financial Enterprises

🚀 Instantly query financial reports, competitor insights, and support tickets using AI-powered natural language.

---

✨ Key Features

• 📂 Google Drive Integration: Automatically syncs and indexes PDFs/financial documents.
• 💬 Voice & Text Chat: Ask questions like “Show me Q3 revenue trends” via voice or text.
• 🔍 Context-Aware Answers: RAG (Retrieval-Augmented Generation) ensures responses are grounded in your documents.
• 📊 User Dashboard: Track query history, manage data sources, and view insights.
• 🔒 Secure Authentication: JWT-based login to protect sensitive financial data.
• ⚡ Real-Time Updates: Detects and processes new/updated files instantly.

---

🛠️ Tech Stack

• Frontend: React, Tailwind CSS, Speech-to-Text API
• Backend: Node.js (Express), Flask (Python)
• Database: MongoDB (user auth/history)
• AI/ML:
  • Mistral-7B (query routing)
  • Gemini 1.5 Pro (answer generation)
  • FAISS (vector search) + SentenceTransformers (embeddings)
• APIs: Google Drive API, CUDA (for GPU acceleration)

---

🚀 Setup Guide

Prerequisites

• Hardware: NVIDIA GPU (RTX recommended) for local ML inference.
• Software:
  • Node.js (v18+), Python (3.9+)
  • CUDA Toolkit (for GPU support)

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/FinanceBrain-AI.git
   cd FinanceBrain-AI

2. Frontend Setup:

   cd frontend
   npm install
   cp .env.example .env  # Add your API keys (Google Drive, Speech-to-Text)
   npm run dev  # Runs on http://localhost:5173

3. Backend Setup:

   cd ../backend
   npm install
   cp .env.example .env  # Add MongoDB URI, JWT secret
   npm start  # Runs on http://localhost:3000

4. Flask (AI Service) Setup:

   cd ../flask-ai-service
   pip install -r requirements.txt
   cp .env.example .env  # Add Gemini/Mistral API keys, Google Drive service account
   python app.py  # Runs on http://localhost:5000

5. Google Drive Sync:

   • Place your Google Drive service account JSON in flask-ai-service/config/.
   • Configure folder IDs in .env (e.g., FINANCIAL_REPORTS_FOLDER_ID=xyz).

---

🐛 Common Issues & Fixes

• CUDA Errors:
  • Ensure NVIDIA drivers are updated.
  • Verify CUDA version compatibility with PyTorch (nvcc --version).
• Google Drive Sync Failures:
  • Check service account permissions.
  • Validate folder IDs in .env.
• Slow Responses:
  • Use quantized Mistral-7B (4-bit) for lower GPU memory usage.

---

📈 Use Case: Financial Enterprises

• Portfolio Managers: Query earnings reports without manual searches.
• Competitor Analysts: Ask “What’s Company X’s market share?” and get sourced answers.
• Support Teams: Resolve client queries using up-to-date policy documents.

---

🔮 Future Improvements

• Add Slack/Microsoft Teams integration.
• Support Excel/PPT files via custom parsers.
• Deploy scalable vector DB (Milvus/Pinecone).

---

📜 License

MIT License. See LICENSE.

---

💬 Contact

Team VAAHM

• Harshvardhan Saini
• Anirban Das
• Mukil M
• Mohd Ashaz Khan
• Vaibhav Aryan
• Hackfest 2025 | IIT (ISM) Dhanbad

---

🎯 Why FinanceBrain?

"Eliminate manual document searches—get AI-powered answers in seconds, grounded in your financial data."

---

✅ Key Notes for Your Repo

• Replace placeholder links (demo GIF, LICENSE, contact).
• Add screenshots in /assets for visual appeal.
• For CUDA issues, link to detailed troubleshooting docs in your wiki.