Multi-Agent AI Document Classifier & Processor

🚀 Smart Document Triage with PDF, JSON, Email Routing 📁 Built with Python, LangChain (LLM), Redis, and Streamlit

---

🔍 Overview

This project is a multi-agent AI system that accepts documents in various formats (PDF, Email, JSON) and intelligently:

• Classifies the document format and business intent (Invoice, RFQ, Complaint, Regulation)
• Routes the input to the correct processing agent (EmailAgent, JSONAgent, PDFAgent)
• Extracts structured data (e.g., sender info, invoice fields, anomalies)
• Stores traceable output in Redis-based shared memory for chain-of-processing transparency

This project leverages LLMs, LangChain, and agent orchestration for real-world intelligent document workflows.

---

🧠 Key Features

Feature	Description
📤 Format Detection	Automatically detects if input is PDF, JSON, or Email
🧾 Intent Classification	Understands if the doc is an Invoice, RFQ, Complaint, or Regulation
🧠 LLM-Powered Extraction	Uses LLMs to extract and clean structured fields from natural text
🔁 Multi-Agent Routing	Orchestrates document to the right processing agent
🧩 Shared Memory via Redis	Persists extracted fields, metadata, thread ID for traceability
📊 Anomaly Detection (JSON)	Flags missing/suspicious fields in structured JSON
📨 CRM Formatting (Email)	Extracts sender info + urgency for downstream CRM integration
📄 Streamlit UI	Upload interface with live logs, previews, and Redis-stored results

---

📁 Folder Structure

.
├── agents/
│   ├── classifier_agent.py
│   ├── email_agent.py
│   ├── json_agent.py
│   └── pdf_agent.py
├── router/orchestrator.py
├── memory/redis_memory.py
├── utils/
│   ├── logger.py
│   ├── json_cleaner.py
│   └── file_handler.py
├── llm/langchain_llm.py
├── app.py  # Streamlit UI
├── requirements.txt
└── README.md

---

🖼️ Sample Output Screenshots

---

📦 Setup Instructions

1. Clone the repository

git clone https://github.com/yourname/multi-agent-ai-docs.git
cd multi-agent-ai-docs

2. Install dependencies

pip install -r requirements.txt

3. Configure Environment Create a .env file:

GROQ_API_KEY=your_groq_api_key
REDIS_HOST=localhost
REDIS_PORT=6379

4. Run the Streamlit App

streamlit run app.py

---

🛠️ Requirements

langchain
langchain-groq
redis
streamlit
python-dotenv
pdfplumber  # for PDF extraction

---

🌐 Project Highlights

• Real-time classification using LLMs (via LangChain)
• Seamless orchestration across multi-modal document types
• Context-sharing using Redis-based memory architecture
• Fully interactive UI with Streamlit
• Highly modular structure for extensibility to other formats or intents

---

Thank you for checking out this project 🙌