IBM EXA Integration

Powered by Exa

Demo Video

This project combines IBM WatsonX AI with Exa's web search to create a smart AI assistant that can search the internet and answer questions with up-to-date information.

What it Does

1. You ask a question
2. The AI (using IBM WatsonX) decides if it needs to search the web
3. If needed, it uses Exa to search the internet
4. The AI then gives you an answer using both the search results and its own knowledge
5. It includes sources and links in its answers

Features

• Uses IBM WatsonX's Mistral model for smart decision making
• Searches the web using Exa for current information
• Automatically cites sources in answers
• Handles both simple questions and complex topics

Setup

1. You need two API keys:

   • IBM WatsonX API key
   • Exa API key

2. Create a .env file with your keys:

IBM_WATSONX_API_KEY=your_ibm_key_here
IBM_WATSONX_URL=https://us-south.ml.cloud.ibm.com
IBM_WATSONX_PROJECT_ID=your_project_id_here
EXA_API_KEY=your_exa_key_here

3. Install Python requirements:

python -m venv venv
source venv/bin/activate  # On Unix/macOS
pip install -r requirements.txt

How to Use

There are several ways to try out this integration:

1. Using Jupyter Notebook: Run the ibm_exa_integration.ipynb notebook to interactively try out the integration in a notebook environment.

2. Running the Example Script:

python -m src.example

3. Using in Your Code:

from src.model import WatsonXModel

# Initialize the model
model = WatsonXModel(exa_api_key="your_exa_key")

# Ask a question
answer = model.ask("What are the latest developments in AI regulation?")
print(answer)

Project Structure

ibm-exa/
├── src/
│   ├── client.py      # IBM WatsonX setup
│   ├── exa_client.py  # Web search client
│   ├── model.py       # Main AI model
│   ├── tools.py       # Search tool definition
│   └── example.py     # Usage example
├── .env              # Your API keys
└── requirements.txt  # Python packages needed

How it Works

1. Tool Calling: The AI can decide when it needs to search the web for information
2. Web Search: Uses Exa's API to find relevant and current information
3. Smart Answers: Combines search results with the AI's knowledge

Requirements

• Python 3.8 or higher
• IBM WatsonX account and API key
• Exa API key