Multi-Agent Report Builder App

Overview

This project is a Streamlit web application that generates structured reports using a multi-stage orchestration workflow. The system leverages LLMs and a modular graph-based approach to plan, parallelize, and synthesize report sections on any user-specified topic.

Features

• Interactive Streamlit Interface: Enter a topic and generate AI-curated multi-section reports in your browser.
• Orchestrator-Worker Design: Tasks are broken into subtasks (sections), assigned to parallel AI workers, and results are synthesized automatically.
• Customizable Topics: Flexible for any topic specified by the user.
• Graph-Based Workflow: Visualizes and manages report generation steps as a directed state graph.

Tech Stack

- Python 3.10+
- Streamlit (UI)
- LangGraph (workflow orchestration)
- LangChain-Groq (LLM integration)
- Pydantic (data modeling)
- python-dotenv (environment variable management)

Installation

To install the MultiAgentReportBuilder project, follow these steps:

1. Clone the repository:

   git clone https://github.com/JSM2512/MultiAgentReportBuilder.git

2. Navigate into the project directory with environment creation:

   # Using conda
   conda create -n venv python=3.12
   conda activate venv/

3. Install the required dependencies:

   pip install -r requirements.txt

Usage with Streamlit

To use the Code with Streamlit:

1. Go to terminal:

   cd Code
   streamlit run app.py

2. This will run the app on your browser.