EventScribe AI 🎯

EventScribe AI is a full-stack, decoupled web application featuring an autonomous AI agent that dynamically queries database to generate rich, context-aware event descriptions. It uses the Model Context Protocol (MCP) to securely bridglive data with large language models This project demonstrates a professional, production-ready architecture where the frontend (Vercel), backend server (Railway)and database (Aiven) are all independently deployed and scalable services Companion Repository: [MySQL-MCP-Server](https://github.com/AbeqqalMarouane/MySQL-MCP-Server

✨ Feature

🤖 Autonomous AI Agent

• Dynamic SQL Generation: The AI agent analyzes user prompts and the database schema to write its own SQL queries on thfly.
• Schema-Aware Context: The agent first fetches the database schema via a secure MCP resource, allowing it to understantable structures and relationships before acting.
• Secure Data Access: The Next.js application has zero direct database credentials. All database interactions arfunneled through the secure, read-only MCP server.
• Smart Prompt Enhancement: The final prompt sent to the LLM is automatically enriched with the precise data the agenchose to fetch

📚 History Management & UI

• Session persistence with local storage.
• Collapsible sidebar to revisit and regenerate previous outputs.
• Full CRUD interface for managing all database entities.
• Professional, responsive UI built with Tailwind CSS and shadcn/ui

🏗️ Production Architectur

This project utilizes a modern, decoupled architecture perfect for scalability and security. The frontend, backend server, andatabase are all separate, independently deployed services that communicate over the internet

graph TD
    A[User's Browser] --> B["EventScribe AI (Next.js App on Vercel)"];
    B -- "API Request to /api/generate" --> B;
    B -- "HTTP Request via MCP" --> C{"MySQL MCP Server on Railway"};
    E["Gemini LLM API"] <--> B;
    C -- "SQL Query" --> D[(MySQL DB on Aiven)];
    D -- "Returns Data" --> C;
    C -- "Returns MCP Response" --> B;

🚀 Reusing the AI Enhancement Service in Other Project

The core logic of this application is available as a reusable service. You can integrate this AI-powered prompt enhancemeninto any other Next.js project with a few simple steps

Step 1: Install Dependencies

In your existing project's terminal, install the necessary libraries:

npm install @google/generative-ai @modelcontextprotocol/sdk

Step 2: Copy the Service File

Copy the lib/eventGenerator.ts file from this project into a lib/ directory in your own project's root

Step 3: Configure Environment Variables

In your project's .env.local file, add the following two variables. The service will automatically use them.

# Your API key for the Google AI model
GEMINI_API_KEY="your_secret_gemini_api_key_here
# The public URL of a deployed MySQL MCP Server
NEXT_PUBLIC_MCP_SERVER_URL="https://your-mysql-mcp-server-production.up.railway.app/mcp"

Step 4: Use the Function in Your Code

Now, in any server-side part of your application (like an API route), you can import and call the `generateEventDescriptionfunction Example Usage:

import { generateEventDescription } from '@/lib/eventGenerator'
async function yourFunction() {
  const simplePrompt = "Tell me about the DevOps conference";
  
  try {
    const result = await generateEventDescription(simplePrompt);
    console.log("Enhanced Description:", result.description);
    console.log("Fetched Context:", result.context);
  } catch (error) {
    console.error("Enhancement failed:", error);
  }
}

📖 Deployment Guide for This Projec

To deploy your own instance of EventScribe AI, you will need to set up the three separate services

Part 1: Database (Aiven)

1. Create a free MySQL database on a cloud provider like Aiven.
2. Get your Service URI. This contains your database credentials.
3. Run the seed script from this project locally (see "Local Development" below) to set up the tables and data in your necloud database

Part 2: Backend - The MCP Server (Railway)

1. Clone the MySQL-MCP-Server repository.
2. Deploy it as a Web Service on Railway.
3. In the Railway project's Variables tab, set the DB_HOST, DB_USER, DB_PASSWORD, etc., using the credentials froyour Aiven database.
4. Railway will provide a public URL. This is your live MCP server endpoint

Part 3: Frontend - The Next.js App (Vercel)

1. Fork and clone this repository.
2. Import the project into Vercel.
3. In the Vercel project's Environment Variables settings, add:
   • GEMINI_API_KEY: Your key from Google AI Studio.
   • NEXT_PUBLIC_MCP_SERVER_URL: The full public URL of your deployed Railway server from Part 2, including the /mcp path.
4. Deploy

💻 Local Developmen

To run this application on your local machine for development

1. Clone this repository and install dependencies (npm install).
2. Set up the .env.local file as described in the "Reusing the AI Service" section, pointing it to your deployed Railwaserver.
3. Run the development server:

   npm run dev
   

4. Open http://localhost:3000 in your browser

🤝 Contributin

Contributions are welcome! Please feel free to fork the repository, make your changes, and submit a pull request

📄 Licens

This project is licensed under the MIT License. See the LICENSE file for details.