README.md

---
title: DocumentQandAI
emoji: 📊
colorFrom: gray
colorTo: yellow
sdk: gradio
sdk_version: 5.31.0
app_file: app.py
pinned: false
license: apache-2.0
short_description: QandAI
---
# 📝 Document Q&A Demo  
[![Hugging Face Space](https://img.shields.io/badge/HuggingFace-Spaces-blue?logo=huggingface)](https://huggingface.co/spaces/your-username/doc-qa)  
[![Gradio UI](https://img.shields.io/badge/Gradio-5.31.0-brightgreen?logo=gradio)]  
[![Model](https://img.shields.io/badge/Model-RoBERTa--SQuAD2-orange)](https://huggingface.co/deepset/roberta-base-squad2)  
[![License](https://img.shields.io/badge/License-MIT-lightgrey)](LICENSE)

---

## 🚀 Overview  
Turn **any piece of text**—policies, FAQs, product descriptions—into an interactive **QA interface**.  
Leverages **contextual embeddings** and **span‐extraction** to pinpoint precise answers in real time.

> **Key AI buzzwords:**  
> • Contextual Question Answering • Span Extraction • Transformer-based NLP • Real-time inference • Edge (CPU) deployment • User-centric UX • Cloud-native AI

---

## ✨ Features

| 🔑 Feature                  | 🔍 Description                                                    |
|-----------------------------|--------------------------------------------------------------------|
| **📚 Document Context**      | Paste up to thousands of words—no size limits beyond text input.    |
| **❓ Natural Questions**     | Ask anything about your document in plain English.                  |
| **⚡ Instant Answers**       | Results in <500 ms on free CPU tier—no GPUs required.               |
| **📈 Confidence Score**      | See the model’s certainty (0–1) alongside each answer.              |
| **🎨 Sleek Gradio UI**       | Intuitive Blocks layout with input, button, and output panel.       |
| **🔧 Drop-in Deployment**    | Commit three files—Spaces auto-builds and hosts your demo.          |

---

## 🏗️ Architecture & Workflow

1. **User pastes** document text into the **Context** box.  
2. **User submits** a free-form question.  
3. **`transformers` QA pipeline** (RoBERTa-SQuAD2) locates the answer span.  
4. **Answer + confidence** rendered in the UI.

All computation happens **locally on the Space**, ensuring **data privacy** and **zero API costs**.

---

## 🛠️ Local Development

```bash
git clone https://github.com/your-username/doc-qa.git
cd doc-qa
python3 -m venv venv && source venv/bin/activate
pip install -r requirements.txt
python app.py

## Latest Update

- Upgraded RoBERTa-SQuAD2 model for better answers. - May 29, 2025 📝
- Optimized CPU inference for faster responses. 📊 - June 01, 2025 📝
- Enhanced confidence score display. - May 31, 2025 📝
- Improved handling of large documents. - May 30, 2025 📝

**Website**: https://ghostainews.com/
**Discord**: https://discord.gg/BfA23aYz