<img width="1024" height="1536" alt="image" src="https://github.com/user-attachments/assets/6c9d8207-11bf-4cc5-88e1-90eb6662ac9a" />

🚀 AstraRAG – Agentic RAG Chatbot

AstraRAG is a production-ready Agentic Retrieval-Augmented Generation (RAG) system that combines multi-agent orchestration, semantic document retrieval, and LLM-powered reasoning to deliver accurate, context-aware responses from custom knowledge bases.

Built using CrewAI, LlamaIndex, ChromaDB, and Groq LLMs, the system leverages autonomous AI agents, vector embeddings, contextual memory, and intelligent retrieval pipelines to provide fast and relevant answers across large document collections.

---

✨ Features

🤖 Multi-Agent Architecture

• Autonomous AI agents powered by CrewAI
• Specialized agents for retrieval, reasoning, and response generation
• Agent collaboration for enhanced answer quality

📚 Retrieval-Augmented Generation (RAG)

• Semantic search using vector embeddings
• Context-aware document retrieval
• Dynamic context injection into LLM prompts
• Support for large document collections

🧠 Contextual Memory

• Conversation-aware interactions
• Maintains context across user sessions
• Improves response consistency and relevance

⚡ High-Performance Inference

• Groq-powered LLM inference
• Sub-300ms end-to-end response latency
• Optimized retrieval and generation pipeline

🌐 Production-Ready Deployment

• FastAPI backend with REST APIs
• Docker containerization
• AWS EC2 deployment support
• Streamlit frontend interface

---

🏗️ System Architecture

                    ┌─────────────────┐
                    │    User Query   │
                    └────────┬────────┘
                             │
                             ▼
                 ┌─────────────────────┐
                 │  Streamlit Frontend │
                 └────────┬────────────┘
                          │
                          ▼
                 ┌─────────────────────┐
                 │   FastAPI Backend   │
                 └────────┬────────────┘
                          │
            ┌─────────────┴─────────────┐
            ▼                           ▼

   ┌────────────────┐         ┌─────────────────┐
   │ CrewAI Agents  │         │ Context Memory  │
   └───────┬────────┘         └────────┬────────┘
           │                           │
           ▼                           ▼

   ┌─────────────────────────────────────┐
   │      Retrieval Pipeline             │
   │  LlamaIndex + ChromaDB Vector Store │
   └────────────────┬────────────────────┘
                    │
                    ▼

          ┌──────────────────────┐
          │ Relevant Documents   │
          └──────────┬───────────┘
                     │
                     ▼

          ┌──────────────────────┐
          │    Groq LLM Engine   │
          └──────────┬───────────┘
                     │
                     ▼

              Final Response

---

🛠️ Tech Stack

AI & LLM

• CrewAI
• Groq
• LlamaIndex

Vector Database

• ChromaDB

Backend

• FastAPI
• Python

Frontend

• Streamlit

DevOps & Deployment

• Docker
• AWS EC2

Data Processing

• Embeddings
• Semantic Search
• Document Chunking
• Contextual Memory

---

📈 Performance Highlights

| Metric | Result | | ------------------------------ | ---------------- | | Indexed Documents | 200+ | | Response Latency | < 300ms | | Retrieval Accuracy Improvement | ~40% | | Deployment | Docker + AWS EC2 | | API Type | RESTful |

---

📂 Project Structure

AstraRAG/
│
├── agents/
│   ├── retrieval_agent.py
│   ├── reasoning_agent.py
│   └── response_agent.py
│
├── backend/
│   ├── main.py
│   ├── routes/
│   └── services/
│
├── frontend/
│   └── streamlit_app.py
│
├── vector_store/
│   └── chromadb/
│
├── documents/
│
├── embeddings/
│
├── docker/
│
├── requirements.txt
├── Dockerfile
├── .env.example
└── README.md

---

⚙️ Installation

Clone Repository

git clone https://github.com/your-username/AstraRAG.git

cd AstraRAG

Create Virtual Environment

python -m venv venv

Activate Environment

Windows:

venv\Scripts\activate

Linux / Mac:

source venv/bin/activate

Install Dependencies

pip install -r requirements.txt

---

🔑 Environment Variables

Create a .env file in the project root:

GROQ_API_KEY=your_groq_api_key

---

▶️ Run FastAPI Backend

uvicorn main:app --reload

Backend will run on:

http://localhost:8000

API Documentation:

http://localhost:8000/docs

---

🎨 Run Streamlit Frontend

streamlit run app.py

Frontend URL:

http://localhost:8501

---

🐳 Docker Deployment

Build Docker Image:

docker build -t astrarag .

Run Container:

docker run -p 8000:8000 astrarag

---

☁️ AWS Deployment

1. Launch AWS EC2 Instance
2. SSH into Instance
3. Install Docker
4. Pull Repository
5. Build Docker Image
6. Run Container
7. Expose Required Ports

Example:

docker build -t astrarag .
docker run -d -p 8000:8000 astrarag

---

🎯 Use Cases

• Enterprise Knowledge Assistants
• Internal Documentation Search
• Customer Support Automation
• Research Assistant Systems
• Educational Knowledge Bases
• AI-Powered Search Applications

---

🔮 Future Enhancements

• Multi-modal document support
• Authentication & user management
• Advanced agent collaboration workflows
• Streaming responses
• Knowledge graph integration
• Kubernetes deployment
• Monitoring and observability dashboards

---

👨‍💻 Author

Arav Chandra

Passionate about AI Engineering, Agentic Systems, LLM Applications, Cloud Deployment, and Building Production-Ready AI Products.

---

⭐ Support

If you found this project useful, consider giving it a star on GitHub. It helps increase visibility and supports future development.