Recipes

Migrate from OpenAI

Moving an existing OpenAI integration to ATP is usually a two-line change: swap the base URL to the Gateway and the API key to a project key. Your request and response shapes stay exactly the same.

Get a project key

Create a project API key in the console. It starts with atp- and inherits its project's allowed models and credits. See Managing API keys.

Swap the base URL and key

Point the OpenAI SDK at the Gateway and use the atp- key. Everything else — messages, tools, streaming — is unchanged.

from openai import OpenAI

# Before: client = OpenAI(api_key="sk-...")
client = OpenAI(base_url="https://api.atptoken.ai/v1", api_key="atp-...")

r = client.chat.completions.create(
    model="<model from GET /v1/models>",
    messages=[{"role": "user", "content": "hi"}],
)
print(r.choices[0].message.content)

Map your model ids

Model ids on ATP come from GET /v1/models, not OpenAI's catalog. List them and update the model field to an id enabled for your project — a model that isn't enabled returns 403.

Test and confirm

Send a request and check the console: it appears in Request logs with input / output tokens, and the credits it used show on the Usage page. See Usage & logs.

Same shapes, one bill

Because the OpenAI wire format passes through unchanged, response parsing in your app doesn't change. Usage across every model is now metered in credits. See How credits work.

← PreviousTeam with budget caps