Introduction

The Nebul Inference API serves open-weight models over an OpenAI-compatible REST API at https://api.inference.nebul.io/v1. If your code talks to OpenAI, point it at Nebul, swap the key and model ID, and it runs: chat, reasoning, vision, embeddings, audio, and image generation, on GPU clusters in EU data centers. Your data never trains the models; see Privacy & Security.

Requests are billed per token in EUR. Every model's input and output price, context window, and capability flags are listed in the Model Catalog.

Make a request now

Create an API key in Nebul AI Studio and export it:

export NEBUL_API_KEY="sk-your-key"

import os

from openai import OpenAI

client = OpenAI(
    api_key=os.environ["NEBUL_API_KEY"],
    base_url="https://api.inference.nebul.io/v1",
)

response = client.chat.completions.create(
    model="zai-org/GLM-5.3",
    messages=[{"role": "user", "content": "Say OK"}],
)

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

You should see the model's answer to your message. The Quick Start covers key setup and the first request in eight languages; the SDKs page has the same setup for Python, TypeScript, JavaScript, Go, Rust, Java, C++, and the shell.

What you can call

Everything speaks OpenAI-shaped JSON, so the official OpenAI SDKs work with a two-line change. /v1/messages is the exception: it speaks the Anthropic Messages format, so the Anthropic SDKs work against it directly. See SDKs for both.

Where to go next

• Quick Start: key setup, first request in your language, full endpoint list
• Examples: streaming, vision, tools, and reasoning code
• Model Catalog: every model with pricing, context windows, and capability flags
• Errors: status codes, response shapes, and retry rules
• API Reference: full OpenAPI specification