WhatsApp AI Assistant

An intelligent, RAG-powered WhatsApp AI assistant built with FastAPI and LLM APIs. Designed specifically to deliver context-aware, domain-specific insurance support for InsureLLM clients.

---

Key Features

• Retrieval-Augmented Generation (RAG): Integrates vector stores to query InsureLLM policy documents and database records for grounded answers.
• WhatsApp Cloud API Integration: Handles real-time incoming messages, multi-turn user conversations, and status verification via webhooks.
• Low-Latency Async Backend: Built with FastAPI and Uvicorn for asynchronous message routing and fast response times.
• Domain Guardrails: Restricts model queries strictly to insurance coverage, contract terms, company policies, and support services.

---

Tech Stack

Component	Technology
Language	Python 3.11+
Backend Framework	FastAPI, Uvicorn
Package Manager	uv / pip
Vector Database	ChromaDB (Vector_DB)
LLM & Embeddings	OpenAI API / NVIDIA NIM / Custom LLM
Messaging Platform	Meta WhatsApp Cloud API

---

Getting Started

Prerequisites

• Python 3.11 or higher
• uv package manager (recommended) or standard pip
• Meta Developer Account with WhatsApp Cloud API access
• API key for your preferred LLM provider

Installation

1. Clone the repository:

git clone https://github.com/shaahmir/WhatsApp-AI-Assistant.git
cd WhatsApp-AI-Assistant

2. Set up the virtual environment and install dependencies: Using uv:

uv venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
uv sync

Using pip:

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install -r requirements.txt

---

Environment Configuration

1. Create a .env file from the template:

cp .env.example .env

2. Add your credentials to .env:

# WhatsApp Cloud API Configuration
PHONE_NUMBER_ID=YOUR_PHONE_NUMBER_ID
ACCESS_TOKEN=YOUR_ACCESS_TOKEN
VERIFY_TOKEN=YOUR_VERIFY_TOKEN

PHONE_NUMBER=YOUR_PHONE_NUMBER
WHATSAPP_BUSINESS_ACCOUNT_ID=YOUR_WA_BUSINESS_ACC

# LLM Provider Configurations 
BASE_URL=LLM_PROVIDER_URL
API_KEY=YOUR_API_KEY

---

Usage

1. Start the FastAPI server:

python -m uvicorn app:app --reload --port 8000

2. Expose the local server for development:

ngrok http 8000

3. Configure Webhook:

• Go to the Meta Developer Portal under WhatsApp Webhook Settings.
• Set Callback URL to https://<your-ngrok-domain>/webhook.
• Provide the VERIFY_TOKEN defined in your .env.
• Subscribe to the messages event.

---

Project Structure

.
├── Vector_DB/        # Vector database index and persistent storage
├── .env.example      # Environment variable template
├── ai.py             # RAG pipeline logic and LLM prompt processing
├── app.py            # FastAPI entry point and web server routes
├── config.py         # App configuration, guardrails, and environment variables
├── webhook.py        # Webhook authorization and event processing
└── whatsapp.py       # Messaging client for sending WhatsApp payloads

---

License

This project is licensed under the MIT License.