Research Agent with LangGraph

This repository contains a research agent built using LangGraph, which combines the power of Claude (via Anthropic) and Tavily search to provide intelligent research capabilities.

Features

• Interactive chat interface
• Integration with Claude 3.5 Sonnet
• Real-time web search using Tavily
• Stateful conversation management using LangGraph

Prerequisites

• Python 3.8 or higher
• pip (Python package manager)
• Virtual environment (recommended)

Setup

1. Clone the repository:

git clone <repository-url>
cd langraph_poc

2. Create and activate a virtual environment:

python -m venv lang_env
source lang_env/bin/activate  # On Windows, use: lang_env\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Set up environment variables: Create a .env file in the root directory with the following content:

ANTHROPIC_API_KEY=your_anthropic_api_key
TAVILY_API_KEY=your_tavily_api_key

You can obtain these API keys from:

• Anthropic API: https://console.anthropic.com/
• Tavily API: https://tavily.com/

Usage

1. Make sure your virtual environment is activated
2. Run the research agent:

python research_agent.py

3. Interact with the agent by typing your questions. The agent will:

   • Process your input
   • Use Claude for understanding and response generation
   • Perform web searches when needed
   • Provide comprehensive answers

4. To exit the conversation, type:

   • quit
   • exit
   • q

Project Structure

• research_agent.py: Main application file containing the LangGraph implementation
• .env: Environment variables configuration
• requirements.txt: Python dependencies

Dependencies

The project uses several key libraries:

• langchain: For LLM integration
• langgraph: For building stateful applications
• anthropic: For Claude API access
• tavily-python: For web search capabilities
• python-dotenv: For environment variable management

Troubleshooting

If you encounter issues with API keys:

1. Verify that your .env file is properly formatted
2. Ensure the API keys are correctly set in your environment
3. Check that you have sufficient API credits/quota

License

[Add your license information here]

Contributing

[Add contribution guidelines here]