Agentic system built with LangGraph, LangChain, Gemini 1.5 Pro, Tavily API, FPDF & Streamlit — automates deep company research, performs intelligent web crawling, and generates structured PDF reports with citations and zero manual effort.

AI-Agent-Based Deep Research

Live Demo

A sophisticated AI-powered research automation system that performs comprehensive company research, generates structured reports, and delivers polished PDF documents with minimal human input.

Overview

This project leverages Large Language Models and specialized AI agents to automate the process of gathering, analyzing, and synthesizing business intelligence. The system follows a structured workflow to transform a simple company name and keywords into a comprehensive, professionally formatted research report.

Key Features

• Automated Research Gathering: Conducts parallel web searches across multiple queries to gather comprehensive company information
• Intelligent Document Curation: Filters and prioritizes the most relevant documents for report generation
• Structured Report Generation: Creates well-organized reports with standardized sections
• Citation Management: Automatically includes citations with quoted text and source URLs
• PDF Export: Converts markdown reports to properly formatted PDF documents
• Rate Limiting: Implements token bucket algorithm to manage API usage efficiently

Architecture

The system is built using a modular architecture based on LangGraph's StateGraph framework:

            ┌─────────┐
            │         │
            ▼         │
┌─────────┐     ┌─────┴───┐
│ research ├─────► tools  │
└────┬────┘     └─────────┘
     │
     ▼
┌─────────┐     ┌─────────┐     ┌─────────┐
│ curate  ├─────► write   ├─────► publish ├─────► END
└─────────┘     └─────────┘     └─────────┘

Core Components

1. Research Node: Coordinates the research process using the Google Generative AI (Gemini)
2. Tools Node: Executes web searches using the Tavily API
3. Curate Node: Filters and processes coll