LLM for Unity Integration

LLM for Unity Integration

A Unity package that connects LLMUnity to Velesio AI Server for seamless LLM integration.

Technical Overview

This integration provides a bridge between Unity and Velesio AI Server’s /completion endpoint:

1

Unity Game → LLMUnity Package → HTTP POST → Velesio AI Server → LLM Response

API Integration

Endpoint: POST /completion

Request Format:

1
2
3
4
5

{
  "prompt": "Your prompt text here",
  "max_tokens": 150,
  "temperature": 0.7
}

Response Format:

1
2
3
4
5
6
7
8

{
  "text": "Generated response text",
  "usage": {
    "prompt_tokens": 10,
    "completion_tokens": 25,
    "total_tokens": 35
  }
}

Unity Implementation

The integration configures LLMUnity to use Velesio AI Server as a remote endpoint instead of running local models:

1
2
3
4
5
6
7
8
9

// Configure LLMUnity to use Velesio AI Server
var llmConfig = new LLMConfiguration
{
    remote = true,
    host = "your-Velesio-server.com",
    port = 8000,
    endpoint = "/completion",
    apiKey = "your-api-token"
};

Authentication

All requests include Bearer token authentication:

1

Authorization: Bearer your-api-token

Benefits

  • No Local Processing: Offloads LLM computation to dedicated GPU servers
  • Scalability: Handle multiple Unity clients with distributed inference
  • Model Flexibility: Switch between different LLM models server-side
  • Performance: Optimized inference on dedicated hardware