🤖 AI Document Q&A System

(Spring Boot 2.7.14 • Java 8 • MySQL • HTML UI)

An AI-powered web application that allows users to upload PDF, Audio, or Video files and interact with them using natural language questions.

The system extracts text from documents, converts speech to text for audio/video files, generates summaries, and answers questions using an AI model.

---

🚀 Features

✅ Upload PDF, Audio, Video files ✅ Automatic text extraction ✅ Speech-to-Text using Vosk (offline) ✅ AI Question Answering ✅ Document Summarization ✅ Timestamp extraction for media ✅ Media streaming support ✅ JWT Authentication ✅ Chat history saved in database

---

🏗 Tech Stack

Backend

• Java 8
• Spring Boot 2.7.14
• Spring Security + JWT
• Hibernate / JPA
• MySQL

Frontend

• HTML
• CSS
• JavaScript

AI / Processing

• Groq / OpenAI API
• Vosk Speech Recognition
• JAVE Audio Converter

---

⚡ Quick Start

1️⃣ Create MySQL Database

CREATE DATABASE Ai_chat;

---

2️⃣ Configure application.yml

Edit:

src/main/resources/application.yml

Update:

spring:
  datasource:
    url: jdbc:mysql://localhost:3306/Ai_chat
    username: root
    password: YOUR_PASSWORD

openai:
  api:
    key: sk-YOUR_API_KEY

---

3️⃣ Run Application

From STS / IntelliJ

Run:

AiChatApplication.java → Run as Spring Boot App

OR terminal

mvn spring-boot:run

---

🌐 Open Application

http://localhost:8080

---

📡 API Endpoints

Authentication (No token)

Method	Endpoint	Body
POST	/api/auth/register	{ "username":"","email":"","password":"" }
POST	/api/auth/login	{ "username":"","password":"" }

---

Documents (Bearer Token Required)

Method	Endpoint	Description
POST	/api/documents/upload	Upload file
GET	/api/documents	List documents
GET	/api/documents/{id}	Get document
GET	/api/documents/{id}/summary	Get summary
GET	/api/documents/{id}/timestamps	Get timestamps
GET	/api/documents/{id}/stream	Stream media
DELETE	/api/documents/{id}	Delete document

---

Chat

Method	Endpoint	Body
POST	/api/chat/ask	{ "documentId":1, "question":"What is this about?" }
GET	/api/chat/history/{documentId}	Chat history

---

📘 Swagger UI

Open:

http://localhost:8080/swagger-ui.html

---

🧪 Run Tests

mvn clean test
mvn jacoco:report

Coverage:

target/site/jacoco/index.html

---

🐳 Docker (Optional)

docker-compose up --build

---

📂 Project Structure

src/main/java/com/aichat
 ├── controller
 ├── service
 ├── repository
 ├── entity
 ├── security

src/main/resources
 ├── static (HTML frontend)
 ├── application.yml

---

🧠 How It Works

1. User uploads file
2. Backend detects file type
3. PDF → text extracted
4. Audio/Video → converted → speech-to-text
5. Transcript saved in MySQL
6. User asks question
7. Transcript + question sent to AI
8. AI response returned

---

📸 Application Screenshots

🔹 Swagger API

🔹 Docker Running

---

👨‍💻 Author

Vishal Pawar Java Full-Stack Developer

---

⭐ Notes

• MySQL must be running
• Vosk model must be placed in project root
• Requires valid AI API key

---

📜 License

For educational / assessment submission.