📄 SmartDocs – AI-Powered PDF Chatbot

SmartDocs is an advanced AI-powered assistant that allows users to interact with PDF documents intelligently. It leverages state-of-the-art Meta Llama 3.1 models (8B, 70B, 405B) to provide accurate and context-aware answers based on the document’s content.

With SmartDocs, you can:
✔️ Upload a PDF and instantly extract its content
✔️ Ask questions and get AI-driven responses based on the document
✔️ Save your queries and responses for later reference

Whether you're a researcher, student, or professional, SmartDocs streamlines the way you engage with PDFs, making information retrieval effortless and efficient.

📹 Watch the Demo Video:

SmartDocs_Preview.mp4

---

✨ Features

✔️ Flexible AI Model Selection – Choose between Meta Llama 3.1 (8B, 70B, 405B)
✔️ PDF Processing – Extracts and chunks text for efficient retrieval
✔️ AI-Powered Query System – Ask questions, get relevant answers instantly
✔️ Vector Database Integration – Uses FAISS for fast & accurate information retrieval
✔️ Query Export – Save your queries and responses as a .txt file

---

🛠️ Tech Stack

• Python
• Libraries:
  • PyPDF2 – Extract text from PDFs
  • faiss – Efficient vector search database
  • sentence_transformers – Embedding text
  • openai – AI model API for answering queries
  • customtkinter – Modern GUI for user interaction

---

📥 Installation

1️⃣ Clone the Repository

git clone https://github.com/Arthur-101/SmartDocs.git
cd SmartDocs

2️⃣ Install Dependencies

python PackageInstaller.py

🚀 Usage

1️⃣ Run SmartDocs

python Main.py

2️⃣ Select AI Model (Optional, default is Llama 405B)

Upon launching, you can choose between Meta Llama 3.1 (8B, 70B, 405B). If no selection is made, the default 405B model will be used.

3️⃣ Upload a PDF

Once uploaded, SmartDocs will:

Extract text from the document

Split the text into smaller, meaningful chunks

Generate embeddings and store them in a FAISS vector database

4️⃣ Query the Document

A query window opens where you can type your questions SmartDocs retrieves the most relevant answers based on the PDF’s content

5️⃣ Export Queries (Optional)

Users can save their queries and responses in .txt format for future reference

🔮 Future Plans

1. Multi-PDF Support – Upload and process multiple PDFs at once

2. Support for More File Formats – DOCX, TXT, and more

3. TTS & STT Integration – Convert text to speech and speech to text

📜 License

This project is open-source under the MIT License