Overview

Serving models efficiently requires optimization across the inference stack. New open-source/open-weight models are released on a frequent cadence, with each novel model architecture requiring specific optimizations for efficient serving. X-Serve provides an inference endpoint to use our optimized inference stack.

Quickstart

from openai import OpenAI

client = OpenAI(
    base_url="https://api.netpreme.com/v1",
    api_key="NETPREME_API_KEY",
)

response = client.chat.completions.create(
    model="moonshotai/Kimi-K3",
    messages=[{"role": "user", "content": "In one sentence, what can you help me with?"}],
)

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

For more details, please refer to the Quickstart.

API

X-serve supports both OpenAI-compatible API and harness integration. For a full list of supported harnesses and integration methods, please refer to the Harness Integration

---

A guide to making inferece requests are shown in Quickstart. Available models are shown in the model catalog. Coding agent integration is shown in Harness Integration. API Documentation are shown here.