Vac

An AI-driven application providing real-time commentary for Esports events.

---

Overview

Vac is an Artificial Intelligence (AI)-powered application that generates commentary for Esports events like CS:GO, Dota 2, and Call of Duty in real-time. It integrates cutting-edge technologies such as GPT-3.5, Go to deliver dynamic, insightful, and engaging commentary. The backend efficiently processes Esports data for generating commentary.

---

Features

• Real-Time Commentary Generation: Generates engaging and contextually relevant commentary during live Esports matches.

---

Technologies Used

• Backend: Go (Golang)
• AI Model: GPT-3.5
• API Communication: RESTful APIs and WebSocket (via Socket.IO)

---

Commentary Output Example

Figure 1: Example of generated commentary output.

---

Architecture

The Vac system consists of three major components:

1. Server1 (Data Preprocessor): Handles data ingestion, filtering, and structuring.
2. Server3 (AI Commentary Generator): Integrates with GPT-3.5 for generating commentary based on processed data.

Flowchart of Data Processing

Figure 2: Data processing flowchart.

System Architecture

Figure 3: System architecture.

---

Prerequisites

Software

• Go 1.20+

Hardware

• Minimum Requirements:
  • Processor: Intel Core i5
  • RAM: 8GB
  • Storage: 256GB SSD
• Recommended:
  • Processor: Intel Core i7 or equivalent
  • RAM: 16GB
  • Storage: 512GB SSD

---

Installation and Setup

1. Clone this repository:

   git clone https://github.com/karkianmol/vac.git
   cd vac

2. Install dependencies for the servers:

   • Server1:

     cd server1
     go mod tidy

   • Server3:

     cd server3
     go mod tidy

3. Configure the environment variables:

   • Create .env files for both servers with necessary configurations:

     OPENAI_API_KEY=<Your OpenAI Key>
     

4. Start the servers:

   • Server1:

     go run server1.go

   • Server3:

     go run server3.go

---

API Endpoints

Server1

• POST /process: Sends processed data to Server3.

Server3

• POST /commentary: Generates and returns AI commentary.

Commentary API Workflow

Figure 4: API workflow for commentary generation.

---

Data Source

The data used for this project is from the Grid Esports Data Jam 2023. You can access the data files here.

---

How It Works

1. Data Ingestion: Server1 preprocesses Esports data in JSONL format, categorizes events, and sends them to Server3.
2. AI Integration: Server3 processes the data using GPT-3.5 and generates context-rich commentary.

---

Future Enhancements

• Audio Commentary: Generate AI-driven voice commentary with emotional tone.
• Game Support Expansion: Adapt functionality for more Esports titles.
• Enhanced Personalization: Provide user-specific commentary settings.

---

License

This project is licensed under the MIT License.

---