PDF Reader Chat with Ollama LLMs

This project lets you chat with your PDF documents using local Large Language Models (LLMs) served by Ollama. Upload a PDF, select an AI model (Llama 3.1, Mistral, GPT-OSS, or Gemma 2), and ask questions about the document in a chat interface. Answers are streamed live and rendered as Markdown for a rich, interactive experience—all on your own machine. Thanks to Deepseek - Best Frontend visuals, ChatGPT - Backend making genius, Copliot - A sew needle that links the beautiful frontend to the backend; that this exists, a work of the best models working on the best prompts.

---

Features

• Upload any PDF and extract its text for context-aware Q&A.
• Choose from multiple open-source LLMs (Llama 3.1, Mistral, GPT-OSS, Gemma 2).
• Chat interface: Ask questions and get answers about your PDF.
• Streaming responses: See the answer appear live as the model generates it.
• Markdown rendering: Answers support formatting, code, and lists.
• Dark mode: User-friendly interface with light/dark theme toggle.
• Privacy-first: All processing is local—your data never leaves your computer.

---

How It Works

1. Frontend (templates/index.html):

   • Lets you select a model, upload a PDF, and chat.
   • Sends your question and selected model to the backend.
   • Streams and renders the model’s answer as Markdown.

2. Backend (first.py):

   • Flask server handles file uploads, chat requests, and serves the frontend.
   • Extracts text from uploaded PDFs using PyPDF2.
   • For each question, sends a prompt (with PDF context) to the selected Ollama model via its HTTP API.
   • Streams the model’s response back to the frontend using Server-Sent Events (SSE).

3. Ollama:

   • Runs locally and serves LLMs via an HTTP API.
   • Handles the actual language model inference.

---

Getting Started (Local Use Only)

Prerequisites

• Python 3.8+
• Ollama installed and running (ollama serve)
• Models downloaded in Ollama (e.g., ollama pull mistral)
• Install Python dependencies:

  pip install flask PyPDF2 requests

Running the App

1. Start Ollama server (in one terminal):

   ollama serve

2. Start Flask backend (in your project directory):

   python PDF_Reader_Chat/first.py

3. Open your browser to http://localhost:5000

---

Why Is This Helpful?

• Research: Quickly extract insights from academic papers, reports, or e-books.
• Productivity: Summarize, search, or analyze large documents without manual reading.
• Privacy: All data stays on your machine—no cloud uploads.
• Flexibility: Choose the LLM that best fits your needs and hardware.

---

File Structure

PDF_Reader_Chat/
├── first.py                # Flask backend
└── templates/
    └── index.html          # Frontend UI

---

Security & Privacy

• All PDF and chat data is processed locally.
• No data is sent to third-party servers.

---

Credits

• Frontend: Deepseek, Copilot
• Backend: ChatGPT, Copilot
• Models: Ollama open-source LLMs

---

Enjoy chatting with your PDFs using local AI!