How to Integrate Kimi K2 with the OpenAI SDK

Moonshot AI's Kimi K2 is a powerful language model that offers high-performance reasoning and extended context capabilities. By utilizing the OpenAI-compatible SDK, developers can integrate Kimi K2 into their existing workflows without rewriting codebase logic for proprietary APIs. This guide demonstrates how to configure your environment to point the standard OpenAI client at the Kimi K2 endpoint.

In this walkthrough, you will learn how to initialize the SDK with custom base URL parameters and execute a standard chat completion request. This approach ensures you can leverage Kimi's specific capabilities while maintaining a vendor-agnostic infrastructure, allowing for easier model swapping and testing in the future.

Prerequisites

• An active Moonshot AI developer account with a valid API key.
• Python 3.8+ installed on your local environment.
• The latest version of the openai Python library installed via pip.
• Basic familiarity with REST API concepts and environment variables.

Steps

1. 1

   Install the OpenAI Python SDK

   First, ensure you have the OpenAI library installed in your project environment. Run 'pip install openai' in your terminal to get the latest package version which supports standard HTTP client overrides.

2. 2

   Configure Environment Variables

   Store your Kimi API key securely by adding it to your environment variables. Avoid hardcoding the API key directly into your scripts to prevent credential exposure in version control systems.

3. 3

   Initialize the OpenAI Client

   Instantiate the OpenAI client by passing the Kimi endpoint as the 'base_url' and your secret key. Set the 'base_url' to 'https://api.select.ax/v1' to route your requests correctly to the Kimi infrastructure.

4. 4

   Execute a Chat Completion Request

   Use the 'client.chat.completions.create' method as you normally would with standard OpenAI models. Ensure you pass 'smart-select' as the 'model' parameter to target the Kimi K2 engine effectively.

5. 5

   Handle the Response Stream

   Parse the returned object to extract the generated text content from the 'choices' array. If you enabled streaming, iterate over the response object using the provided generator to handle real-time output.

Code

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ.get("KIMI_API_KEY"),
    base_url="https://api.select.ax/v1"
)

response = client.chat.completions.create(
    model="smart-select",
    messages=[{"role": "user", "content": "Explain the benefits of Kimi K2."}]
)

print(response.choices[0].message.content)

Pro tips

Visual guide

Route your models intelligently

Use one API key for routing, fallback, and cost control across model providers.