DOCS / QUICKSTART
Quickstart
Four steps from your first key to a priced request. Use your own provider keys.
/01
Create a key
Create an Untokened key in your dashboard. Add your provider key to your vault, then choose the provider for your product. Keep both keys out of your source code.
Account setup is not available in this preview yet.
/02
Change base_url
Use the OpenAI client you already have. Set its base URL to the Untokened gateway and use your Untokened key.
PYTHON / CLIENT
from openai import OpenAI
client = OpenAI(
api_key="YOUR_UNTOKENED_KEY",
base_url="https://api.untokened.com/v1",
)/03
Send a first request
Send a chat completion through the client. The gateway routes it to your configured upstream provider.
PYTHON / REQUEST
response = client.chat.completions.create(
model="gpt-4.1-mini",
messages=[
{"role": "user", "content": "Hello"}
],
)
print(response.choices[0].message.content)/04
See it in the dashboard
Open the request log in your dashboard. Find the model, token count, cost and timestamp on one row. Add a product or tag to split the numbers by use case.
Next: Gateway