Quickstart

Get your first Meridian API call running in under two minutes.

0

Decide which model to use

Meridian serves the fastest inference across every major provider. Start with one of these low-latency models:

azure-swc/Llama-4-Maverick
Fastest overall · 0.77s
azure-swc/gpt-4.1
OpenAI-compatible · 0.86s
gpt-5-codex
Bare-metal codex · 1.3s
1

Get your API key

Sign up at meridian.sh/sign-up and grab your key from the API Keys dashboard.

Your key will look like
sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
2

Set environment variables

Export these in your shell or add them to your .env file.

MERIDIAN_API_KEY=sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
MERIDIAN_BASE_URL=https://api.meridian.sh/v1
3

Choose your language

Meridian is OpenAI-compatible — use any standard SDK.

4

Install the SDK

pip install openai
5

Make your first call

Copy this example, replace the API key, and run it.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.meridian.sh/v1",
    api_key="sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
)

response = client.chat.completions.create(
    model="azure-swc/Llama-4-Maverick",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Explain quantum computing in one sentence."},
    ],
    max_tokens=128,
)

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