AI Finance Chatbot for WhatsApp

An AI-powered WhatsApp chatbot that answers finance-related queries using OpenAI's function calling and tool integrations. Built with FastAPI, Twilio, LangChain-style prompt workflows, and PostgreSQL — deployed on Render.com.

🧠 Features

• ✅ Answers finance-related questions via WhatsApp using OpenAI responses
• ✅ Detects intent (stock/market queries) and triggers real-time web search tool
• ✅ Stores full chat history in PostgreSQL via SQLAlchemy ORM
• ✅ Handles multi-part message delivery with auto-chunking
• ✅ Formats AI responses for clean, readable WhatsApp rendering
• ✅ Runs on FastAPI with Twilio WhatsApp integration
• ✅ Logging, error handling, retry logic, and user state tracking via response_id
• ✅ Fully deployed and live on Render (API + DB)

🛠️ Tech Stack

• FastAPI – Backend API framework
• OpenAI SDK – Language + tool-calling responses
• Twilio SDK – WhatsApp integration
• SQLAlchemy + PostgreSQL – Persistent message logging
• Alembic – DB migrations
• Render.com – Prod deployment for API + DB
• python-decouple – Secure environment configuration

⚙️ Installation

🧱 Prerequisites

• Python 3.9+
• A Twilio account with WhatsApp sandbox access
• An OpenAI API key
• A PostgreSQL instance (local or cloud-hosted)
• (Optional) Ngrok for local webhook testing

🧪 Local Setup

1. Clone the repo

git clone https://github.com/nawazkhan/ai-finance-chatbot.git
cd ai-finance-chatbot

2. Create a .env file

# OpenAI
OPENAI_API_KEY=your-openai-key
OPENAI_MODEL=text-completion-model-id

# Twilio
TWILIO_ACCOUNT_SID=your-sid
TWILIO_AUTH_TOKEN=your-auth-token
TWILIO_NUMBER=

# PostgreSQL
DB_DRIVER=postgresql
DB_USER=your_user
DB_PASSWORD=your_password
DB_HOST=localhost
DB_PORT=5432
DB_NAME=your_db

3. Install dependencies

pip install -r requirements.txt

4. Run database migrations

alembic upgrade head

5. Start the server

uvicorn main:app --reload

6. Expose locally with ngrok (for Twilio)

ngrok http 8000

7. Configure Twilio webhook to:

https://your-ngrok-url.ngrok.io/message

🚀 Production Deployment

Deployed on Render.com with:

• FastAPI backend on a Render web service
• PostgreSQL DB as a managed Render database
• Auto-deploy from GitHub + environment variables configured via dashboard

📬 Contact

Built by Nawaz Khan
Email: [email protected]

🛡️ License

MIT License – open for use and modification with attribution