MultiAgentic RAG

This repository showcases the implementation of a Multi-Agent Research RAG (Retriever-Augmented Generation) Tool built with LangGraph. This project leverages the capabilities of agent-based frameworks to handle complex queries by breaking them down into manageable steps, dynamically utilizing tools, and ensuring response accuracy through error correction and hallucination checks.

Getting Started

To get started with this project, follow these steps:

First, clone the repository to your local machine:

git clone https://github.com/nicoladisabato/MultiAgenticRAG.git
cd MultiAgenticRAG

pip install -r requirements.txt

Then open the config.yml file located in the root directory of the project. Set the value of load_documents to true to ensure the necessary documents are loaded into the vector database:

Then run:

python3 -m retriever.retriever

Once the PDF has been processed and indexed, you can start the application by running the following command:

python3 app.py

Now ask your question based on the document: https://sustainability.google/reports/google-2024-environmental-report/