---
title: Chat With Documents
emoji: 🚀
colorFrom: purple
colorTo: purple
sdk: streamlit
sdk_version: 1.13.0
app_file: app.py
pinned: false
---



---

# Chat With Documents 🤖📄

Welcome to the **Chat with Documents** app! 🚀 This Streamlit app allows you to upload PDF and PPT files, extract their content, store the extracted text in a vector store, and interact with it using natural language queries! 🤖💬

Built with **LangChain**, **OpenAI**, **Streamlit**, and **Astra DB**, this project leverages the power of LLMs (Large Language Models) to allow users to chat with their documents like never before. 🧠

---

### 🚀 **Features**

- **PDF & PPT Extraction**: Upload PDF and PowerPoint files to extract text! 📄➡️📝
- **Vector Store**: Automatically stores extracted text in a **Cassandra** vector store. 🔍📚
- **Ask Anything**: Ask questions about the document and get answers powered by **OpenAI**! 🤖❓

---

### 🛠️ **Tech Stack**
- **Streamlit**: Frontend framework to interact with the app.
- **LangChain**: For seamless document processing and querying.
- **OpenAI**: For LLM integration to provide intelligent responses.
- **Astra DB**: Database for storing and managing vectorized text data.
- **Python Libraries**: PyPDF2, python-pptx, cassio, and more.

---



### 💡 **How It Works**

- Upload a **PDF** or **PPT** file using the file uploader. 📤
- The app will extract text from the file using **PyPDF2** (for PDFs) or **python-pptx** (for PPTs). 📄➡️📝
- The extracted text is split into manageable chunks using **LangChain's CharacterTextSplitter**. ✂️
- The chunks are then added to **Cassandra** as vectorized data using **OpenAI embeddings**. 🔄
- Ask any query about the content of your document, and the app will respond using the power of **OpenAI**! 🤖💬

---

### 🎯 **Why Use This?**

- **Make documents interactive**: Easily explore the content of your documents by asking questions.
- **Quick retrieval**: With the text stored in a vector store, you can query the content efficiently.


---

### ✨ **Enjoy the App!** ✨  
Now, go ahead and chat with your documents! 😄

---