🚀 Aakaar AI Agentic System

A high-performance, multi-agent AI system designed for advanced RAG, text summarization, email drafting, and Excel validation. Powered by Groq (primary) and Google Gemini (fallback).

---

🏗️ System Architecture

The project follows a Supervisor-Agent pattern:

• Supervisor Agent: The "brain" that routes user requests to the appropriate specialized agent based on natural language intent.
• RAG Agent: Handles PDF ingestion and context-aware retrieval-augmented generation.
• Summarization Agent: Provides concise summaries of long documents or text snippets.
• Email Agent: Drafts professional emails based on context and user instructions.
• Excel Checker Agent: Validates spreadsheet structure for business report compatibility.

---

⚡ Tech Stack

• Backend: FastAPI (Python 3.11+)
• Frontend: Streamlit (Premium UI with Glassmorphism)
• LLM Engine: Groq SDK (llama-3.1-70b-versatile)
• Fallback: Google Generative AI (gemini-1.5-flash)
• Deployment: Render-ready with dedicated Procfile

---

🛠️ Setup & Installation

1. Prerequisite Keys

You will need API keys from:

1. Groq Cloud
2. Google AI Studio

2. Local Installation

# Clone the repository
git clone https://github.com/Saviour5538/testmultiagent.git
cd testmultiagent

# Install dependencies
pip install -r requirements.txt

# Configure Environment
cp env.template .env
# Edit .env and paste your API keys

3. Running Locally

# Start the Backend (FastAPI)
python -m backend.main

# Start the Frontend (Streamlit)
streamlit run frontend/app.py

---

📂 API Reference (For Platforms & Testing)

The system provides a unified gateway for agent interaction.

Unified Agent Chat

• Endpoint: POST /chat
• Method: POST
• Body (Form Data):
  • message: (string) Your task or query.
  • file_path: (string, optional) Path to an ingested PDF or Excel file.

Knowledge Ingestion

• Endpoint: POST /ingest
• Method: POST
• Body (File): file (PDF binary)
• Returns: file_path for use in context-aware chats.

Excel Validation

• Endpoint: POST /upload-excel
• Method: POST
• Body (File): file (Excel binary)

---

☁️ Deployment

This project is optimized for Render.

1. New Web Service: Connect your GitHub repo.
2. Build Command: pip install -r requirements.txt
3. Start Command: python -m backend.main
4. Environment Variables:
   • GOOGLE_API_KEY: [Your Key]
   • GROQ_API_KEY: [Your Key]
   • PYTHONPATH: .

---

🔒 Security

Sensitive files such as .env and __pycache__ are strictly excluded via .gitignore. An env.template is provided for safe configuration reference.