Skip to main content
Use the OpenAI-compatible base URL with a Tokenless API key: 
https://api.usetokenless.com/openai/v1
Tokenless preserves the OpenAI request and response formats for the routes below, while choosing the upstream model behind a routing mode such as tokenless-pro.

​
Use an OpenAI client

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://api.usetokenless.com/openai/v1",
    api_key=os.environ["TOKENLESS_API_KEY"],
)

response = client.chat.completions.create(
    model="tokenless-pro",
    messages=[{"role": "user", "content": "Hello"}],
)

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

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.usetokenless.com/openai/v1",
  apiKey: process.env.TOKENLESS_API_KEY,
});

const response = await client.chat.completions.create({
  model: "tokenless-pro",
  messages: [{ role: "user", content: "Hello" }],
});

console.log(response.choices[0]?.message.content);

curl https://api.usetokenless.com/openai/v1/chat/completions \
  -H "Authorization: Bearer $TOKENLESS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tokenless-pro",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

​
Use Codex with tkl

With the Tokenless CLI installed, sign in once and launch Codex through the OpenAI-compatible route. No API key setup is required. 
tkl login
tkl codex
For the full upstream specification, see the OpenAI API reference. Use these pages for the Tokenless-compatible routes and routing behavior.