🧠 DocuMind — AI-Powered Document Analyzer

Upload any legal contract or research paper — ask questions in plain English — get accurate answers with exact page citations. No hallucination guaranteed.

🚀 Live Demo

👉 Try DocuMind Live — docu-mind03.streamlit.app

⚡ No installation needed — upload any PDF and start asking questions instantly!

🔌 API (Backend)

FastAPI backend is also deployed separately on Render:

• Base URL: https://documind-api-y0hc.onrender.com
• Swagger Docs: https://documind-api-y0hc.onrender.com/docs

Note: Render free tier may have cold start delay of 30-60 seconds.

---

🎯 What Problem Does This Solve?

Reading a 100-page legal contract or research paper to find one specific clause is time-consuming and error-prone.

DocuMind lets you:

• Upload any PDF document
• Ask questions in plain English
• Get precise answers with page number citations
• Know exactly which part of the document was used
• Never get hallucinated answers — strict document-only responses

---

✨ Features

Feature	Description
📄 PDF Upload	Upload any legal or research PDF
🧩 Smart Chunking	1000-token chunks with 200-token overlap
🔍 Semantic Search	HuggingFace embeddings + ChromaDB vector search
🤖 LLM Answer	Groq LLaMA3-70b generates accurate answers
📌 Source Citations	Every answer shows exact page numbers
📋 Auto Summary	Document summary generated on upload
🟢🔴 Confidence Score	Green = found in doc, Red = not found
🚫 No Hallucination	Strict prompting — only answers from document

---

🏗️ Architecture

User uploads PDF
      ↓
PDF → Pages → Chunks (1000 tokens, 200 overlap)
      ↓
HuggingFace Embeddings (all-MiniLM-L6-v2)
      ↓
ChromaDB Vector Store
      ↓
User asks question → Similarity Search → Top 3 chunks
      ↓
Groq LLaMA3-70b → Answer with page citations
      ↓
Streamlit UI displays answer + confidence + sources

---

🛠️ Tech Stack

Layer	Technology
Frontend	Streamlit
Backend	FastAPI + Uvicorn
RAG Pipeline	LangChain
Embeddings	HuggingFace all-MiniLM-L6-v2
Vector DB	ChromaDB (Persistent)
LLM	Groq API — LLaMA3-70b-versatile
PDF Processing	PyPDF

---

📁 Project Structure

DocuMind/
├── backend/
│   ├── main.py          # FastAPI server — 3 endpoints
│   ├── rag_pipeline.py  # Embeddings + ChromaDB + Groq LLM
│   ├── summarizer.py    # Auto document summary
│   └── utils.py         # PDF loader + chunker
├── frontend/
│   └── app.py           # Streamlit UI
├── uploads/             # Uploaded PDFs stored here
├── chroma_db/           # Vector embeddings stored here
├── .env.example         # Environment variables template
├── .gitignore
└── requirements.txt

---

⚙️ How to Run Locally

1. Clone the repository

git clone https://github.com/bhatt-aditya03/DocuMind.git
cd DocuMind

2. Create virtual environment

python3 -m venv venv
source venv/bin/activate  # Mac/Linux
venv\Scripts\activate     # Windows

3. Install dependencies

pip install -r requirements.txt

4. Set up environment variables

cp .env.example .env
# Add your Groq API key in .env

5. Run the app

streamlit run frontend/app.py

6. Open in browser

http://localhost:8501

---

🔑 Environment Variables

Create a .env file in root directory:

GROQ_API_KEY=your_groq_api_key_here

Get your free Groq API key at console.groq.com

---

📡 API Endpoints (Local Development)

Method	Endpoint	Description
GET	/	API health check
POST	/upload	Upload and process PDF
POST	/ask	Ask question about document
GET	/summary/{doc_id}	Get document summary

Example Request

# Upload PDF
curl -X POST "http://localhost:8000/upload" \
  -F "file=@document.pdf"

# Ask question
curl -X POST "http://localhost:8000/ask" \
  -H "Content-Type: application/json" \
  -d '{"question": "What are the key clauses?", "doc_id": "document"}'

---

💡 Use Cases

• 📜 Legal Contracts — Find specific clauses instantly
• 📚 Research Papers — Extract key findings and methodology
• 📋 Policy Documents — Understand complex policies quickly
• 🎓 Academic Syllabi — Navigate course content easily
• 📑 Business Reports — Quick insights from long reports

---

🗓️ Built In 5 Days

Day	What was built
Day 1	Project setup, dependencies, environment
Day 2	RAG pipeline — PDF chunking, ChromaDB, Groq LLM
Day 3	FastAPI backend — upload, ask, summary endpoints
Day 4	Streamlit UI — chat interface, confidence indicator
Day 5	README, deployment, live on Streamlit Cloud

Post-launch refactor: @lru_cache for embedding model, specific exception handling, shared sanitize_doc_id() utility, temp file cleanup, and English docstrings throughout.

---

🧑‍💻 Author

Aditya Bhatt

---

📄 License

MIT License — free to use, modify and distribute!