Industrial Knowledge Intelligence Assistant

An AI-powered Retrieval-Augmented Generation (RAG) application that allows users to upload PDF documents and ask questions based on their contents. The application retrieves relevant document chunks using vector embeddings and generates accurate responses using Large Language Models.

---

Features

• Upload one or multiple PDF documents
• Automatic document chunking and embedding
• Vector database for semantic retrieval
• AI-powered question answering using Groq LLM
• Multiple AI provider support (Groq, Gemini, OpenAI, Anthropic)
• Modern React frontend
• Beautiful chat interface
• Markdown rendering with syntax highlighting
• Chat history
• Export chat as PDF
• Dark UI
• Responsive design
• Local storage for conversations
• Document source retrieval

---

Tech Stack

Frontend

• React
• Vite
• Tailwind CSS
• Axios
• Framer Motion
• React Markdown
• Hero Icons
• React Hot Toast

Backend

• Flask
• Flask-CORS
• LangChain
• LangChain Groq
• ChromaDB
• Sentence Transformers
• PyPDF Loader

---

Project Structure

ET2/
│
├── backend/
│   ├── app.py
│   ├── requirements.txt
│   ├── uploads/
│   ├── vectorstore/
│   ├── src/
│   │
│   ├── ingest.py
│   ├── rag_chain.py
│   ├── retriever.py
│   └── ...
│
├── frontend/
│   ├── src/
│   ├── public/
│   ├── package.json
│   └── vite.config.js
│
└── README.md

---

System Architecture

          User
            │
            ▼
      React Frontend
            │
            ▼
      Flask REST API
            │
    ┌───────┴────────┐
    ▼                ▼
PDF Upload      Chat Request
    │                │
    ▼                ▼
Chunk Documents   Retrieve Context
    │                │
    ▼                ▼
Vector Store ---> Groq LLM
            │
            ▼
       AI Response
            │
            ▼
        React UI

---

Installation

Clone Repository

git clone https://https://github.com/vedddev/Industary_brain.git

---

Backend Setup

Create virtual environment

python -m venv brain

Activate environment

Windows

brain\Scripts\activate

Linux/Mac

source brain/bin/activate

Install dependencies

pip install -r requirements.txt

Create .env

GROQ_API_KEY=your_api_key

Run backend

python app.py

Backend runs on

https://industary-brain.onrender.com

---

Frontend Setup

Navigate to frontend

cd frontend

Install dependencies

npm install

Run

npm run dev

Frontend runs on

https://industary-brain.vercel.app/

---

API Endpoints

Home

GET /

Response

{
  "message":"Industrial Knowledge Intelligence API"
}

---

Upload PDFs

POST /upload

Form Data

files

Response

{
  "status":"success",
  "documents":2,
  "chunks":164
}

---

Chat

POST /chat

Body

{
    "question":"Explain tensile strength",
    "provider":"groq",
    "api_key":"YOUR_API_KEY"
}

Response

{
    "question":"Explain tensile strength",
    "answer":"...",
    "sources":[]
}

---

Health

GET /health

Response

{
    "status":"healthy"
}

---

Workflow

1. Upload PDF documents.
2. Backend extracts text.
3. Documents are split into chunks.
4. Chunks are converted into embeddings.
5. Embeddings are stored in Chroma Vector Database.
6. User asks a question.
7. Relevant chunks are retrieved.
8. Retrieved context is sent to the LLM.
9. AI generates an answer.
10. Response is displayed in the chat interface.

---

AI Providers

The application is designed to support multiple providers.

• Groq
• Google Gemini
• OpenAI
• Anthropic

Currently implemented:

• ✅ Groq

Future support:

• Gemini
• OpenAI
• Anthropic

---

Frontend Features

• Responsive sidebar
• Chat history
• Markdown rendering
• Code syntax highlighting
• Copy response
• Like/Dislike responses
• Regenerate answer
• Upload PDFs
• Voice input placeholder
• Export chat to PDF
• Animated interface
• Mobile responsive

---

Backend Features

• REST API
• PDF ingestion
• Vector embeddings
• ChromaDB integration
• LangChain retriever
• Retrieval-Augmented Generation (RAG)
• Source references
• Multi-file upload

---

Environment Variables

GROQ_API_KEY=xxxxxxxxxxxxxxxx

---

Dependencies

Backend

• Flask
• Flask-CORS
• LangChain
• LangChain-Groq
• ChromaDB
• Sentence Transformers
• PyPDF

Frontend

• React
• Vite
• Tailwind CSS
• Axios
• Framer Motion
• React Markdown
• Hero Icons

---

Screenshots

Add screenshots here.

Example

screenshots/

home.png

chat.png

upload.png

---

Future Improvements

• Authentication
• Database support
• Conversation memory
• Streaming responses
• OCR for scanned PDFs
• Image understanding
• Audio transcription
• Citation highlighting
• Multi-user support
• Cloud deployment
• Admin dashboard

---

Deployment

Frontend

• Vercel
• Netlify

Backend

• Render
• Railway
• Azure
• AWS

---

Security

• API keys stored locally
• Environment variables ignored using .gitignore
• CORS enabled
• Sensitive files excluded from Git

---

License

This project is licensed under the MIT License.

---

Author

Your Name

Industrial Knowledge Intelligence Assistant

Built with ❤️ using React, Flask, LangChain, ChromaDB, and Groq.