Portkey - Nebius Token Factory documentation

You can work with Nebius Token Factory by using Portkey. The integration is based on the AI Gateway by Portkey. The AI Gateway allows users to send requests to different AI providers and still use the same Portkey interface. The solution assumes that you send requests to the AI Gateway, and then it routes and sends them to Nebius Token Factory. As a result, you interact with Nebius Token Factory models via Portkey. The integration is based on the OpenAI Node.js package and the Nebius Token Factory provider for Portkey. To use the integration, do the following:

Create a Nebius API key for authentication.
Create a Portkey API key for authentication.

Save the keys to the environment variables:

export NEBIUS_API_KEY=<Nebius_API_key>
export PORTKEY_API_KEY=<Portkey_API_key>

Configure the integration:

Node.js
cURL

Install the Portkey SDK:
```
npm i --save portkey-ai
```

Create the following portkey.js script:

import OpenAI from 'openai'; // v4 SDK
import { PORTKEY_GATEWAY_URL, createHeaders } from 'portkey-ai'

const openai = new OpenAI({
  apiKey: process.env["NEBIUS_API_KEY"],
  baseURL: PORTKEY_GATEWAY_URL,
  defaultHeaders: createHeaders({
    provider: "nebius",
    apiKey: process.env["PORTKEY_API_KEY"]
  })
});

async function main() {
  const chatCompletion = await openai.chat.completions.create({
    messages: [{ role: 'user', content: 'Say this is a test' }],
    model: 'deepseek-ai/DeepSeek-R1-fast',
  });

  console.log(chatCompletion.choices);
}

main();

This script sends a multi-message request to the specified model in Nebius Token Factory.

Run the script:

node portkey.js

The output is the following:

[
  {
    finish_reason: 'stop',
    index: 0,
    logprobs: null,
    message: {
      content: 'This is a test. How can I assist you further?',
      refusal: null,
      role: 'assistant',
      audio: null,
      function_call: null,
      tool_calls: []
    },
    stop_reason: null
  }
]

Send the following request:

curl https://api.portkey.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $NEBIUS_API_KEY" \
  -H "x-portkey-api-key: $PORTKEY_API_KEY" \
  -H "x-portkey-provider: nebius" \
  -d '{
    "messages": [{
        "role": "system",
        "content": "You are a helpful assistant."
      },{
        "role": "user",
        "content": "Hello!"
    }]
  }'

The output is the following (below is a pretty-printed version, the curl command returns output in one line):

{
  "id": "chatcmpl-b4586***",
  "choices": [
    {
      "finish_reason": "stop",
      "index": 0,
      "logprobs": null,
      "message": {
        "content": "Hello! How can I assist you today? Feel free to ask any questions or let me know if you need help with anything. 😊",
        "refusal": null,
        "role": "assistant",
        "audio": null,
        "function_call": null,
        "tool_calls": [],
        "reasoning_content": null
      },
      "stop_reason": null
    }
  ],
  "created": 1740159035,
  "model": "Qwen/Qwen2.5-72B-Instruct-fast",
  "object": "chat.completion",
  "service_tier": null,
  "system_fingerprint": null,
  "usage": {
    "completion_tokens": 29,
    "prompt_tokens": 21,
    "total_tokens": 50,
    "completion_tokens_details": null,
    "prompt_tokens_details": null
  },
  "prompt_logprobs": null,
  "provider": "nebius"
}