Usage

Getting started with nilAI is straightforward and involves just two steps:

1. Query the /v1/models endpoint using your preferred programming language or client.
2. Select one of the available models and use its name to query the /v1/chat/completions endpoint.

Since nilAI offers OpenAI API compatibility, you can seamlessly use libraries like openai to interact with the service.

---

All of these endpoints are available for testing in our API section.

• Chat Completion: Standard Chat completion endpoint to generate a response from the AI model
• Attestation: Generate a cryptographic attestation report.
• Health: Check the health status of the NilAI operational status
• Models: Get the list of available models running
• Usage: Track the current usage of tokens with your account

Here is an Python/Node examples for querying the Llama-3.1-8B model:

Python

from openai import OpenAI

# Initialize the OpenAI client

client = OpenAI( # Replace <node> with the specific node identifier
    base_url="https://nilai-<node>.nillion.network/v1/",
    api_key="YOUR_API_KEY"
)

# Send a chat completion request

response = client.chat.completions.create(
    model="meta-llama/Llama-3.1-8B-Instruct",
    messages=[
        {
            "role": "system",
            "content": "You are a helpful assistant."
        },
        {
            "role": "user",
            "content": "What is your name?"
        }
    ],
    stream=False
)

Node

const OpenAI = require('openai');

// Initialize the OpenAI client
const client = new OpenAI({
  baseURL: 'https://nilai-<node>.nillion.network/v1/',
  apiKey: 'YOUR_API_KEY'
});

// Send a chat completion request
async function getChatCompletion() {
  try {
    const response = await client.chat.completions.create({
      model: 'meta-llama/Llama-3.1-8B-Instruct',
      messages: [
        {
          role: 'system',
          content: 'You are a helpful assistant.'
        },
        {
          role: 'user',
          content: 'What is your name?'
        }
      ],
      stream: false
    });

    console.log(response);
  } catch (error) {
    console.error('Error:', error);
  }
}

// Call the function
getChatCompletion();