An AI-powered fintech customer support chatbot built using Node.js, Express, MongoDB, and Groq LLM (LLaMA model).
It simulates a real banking support system that answers customer queries safely using AI + rules, without giving financial advice.
FinBot is a SaaS platform that allows fintech businesses — banks, payment apps, and investment platforms — to instantly create an AI-powered customer support chatbot from their FAQ data.
Instead of spending months building a custom support system, businesses paste their FAQ, get a shareable chat link in seconds, and manage all customer queries from a dedicated admin dashboard.
- AI answers customer questions from business FAQ only
- Quick question chips for common fintech queries
- Typing indicator for better UX
- 👍👎 feedback on every response
- Max 300 character input with live counter
- Sensitive data masking (card numbers → XXXX)
- Personal info detection with safe redirect
- Duplicate unresolved query prevention
- Separate dashboard per business via shareLink
- Total chats, 👍 👎 counts, satisfaction rate
- Top 5 most asked questions
- Unresolved queries list with manual reply
- Auto refresh every 30 seconds
- Mark queries as resolved
- Sensitive number masking (4+ digits → XXXX)
- Personal info detection
- Input validation and sanitization
- Compliance disclaimer on every chat
- No financial advice given by AI
- Fraud queries escalated to human support
- Multi-tenant — each business gets isolated chatbot + dashboard
- Shareable chat link for customers
- Private dashboard link for admin
- Unresolved query escalation system
- Admin reply saved and reused for future same questions
| Layer | Technology |
|---|---|
| Frontend | React.js, Axios, React Router |
| Backend | Node.js, Express.js |
| Database | MongoDB Atlas |
| AI | Groq API (Llama 3.3 70B) |
| Deployment | Vercel (Frontend), Render (Backend) |
Anyone can run this chatbot locally in a few simple steps:
1.Clone the Project
First, download the project to your system:
git clone https://github.com/Pragzh/faq-chatbot.git
cd faq-chatbot
2. Setup Backend
Go into backend folder:
cd backend
npm install
3️. Add Environment Variables
Create a .env file inside the backend folder:
PORT=5000
MONGO_URI=your_mongodb_connection_string
GROQ_API_KEY=your_groq_api_key
👉 You can get:
MongoDB: https://www.mongodb.com/atlas
Groq API key: https://console.groq.com
4️. Start Backend Server
node index.js
If successful, you will see:
Server running on http://localhost:5000
5️. Setup Frontend
Open a new terminal:
cd frontend
npm install
npm start
Frontend will run at:
http://localhost:3000
6️. Use the Chatbot
Now open your browser:
👉 Go to http://localhost:3000
You can:
Create a chatbot instance
Get a shareable link
Send messages
Get AI responses instantly
---
## 📁 Project Structure
faq-chatbot/
├── backend/
│ ├── index.js # All API routes
│ ├── db.js # MongoDB connection
│ ├── package.json
│ └── .env # Environment variables (not committed)
│
└── frontend/
└── src/
├── pages/
│ ├── Setup.jsx # Create chatbot page
│ ├── ChatPage.jsx # Customer chat UI
│ └── Dashboard.jsx # Admin dashboard
└── App.jsx