OpenAI-Compatible API

Tokeness exposes an OpenAI-compatible API. Most SDKs, chat clients, and automation tools that support a custom OpenAI Base URL can connect to Tokeness.

Configuration

Base URL: https://n.tokeness.io/v1
API Key: your Tokeness API key

Install the SDK

npm install openai

pip install openai

Node.js

Create test-tokeness.mjs and replace YOUR_TOKENESS_API_KEY and YOUR_MODEL_NAME with your own values.

import OpenAI from 'openai'

const client = new OpenAI({
  apiKey: 'YOUR_TOKENESS_API_KEY',
  baseURL: 'https://n.tokeness.io/v1'
})

const completion = await client.chat.completions.create({
  model: 'YOUR_MODEL_NAME',
  messages: [
    { role: 'system', content: 'You are a concise assistant.' },
    { role: 'user', content: 'Say hello from Tokeness.' }
  ]
})

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

Run:

node test-tokeness.mjs

Python

Create test_tokeness.py and replace YOUR_TOKENESS_API_KEY and YOUR_MODEL_NAME with your own values.

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_TOKENESS_API_KEY",
    base_url="https://n.tokeness.io/v1",
)

response = client.chat.completions.create(
    model="YOUR_MODEL_NAME",
    messages=[
        {"role": "user", "content": "Say hello from Tokeness."}
    ],
)

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

cURL

curl https://n.tokeness.io/v1/chat/completions \
  -H "Authorization: Bearer $TOKENESS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "YOUR_MODEL_NAME",
    "messages": [
      { "role": "user", "content": "Say hello from Tokeness." }
    ]
  }'

Model names

Model names come from the model marketplace in the Tokeness console or from the API response. When copying a model name, keep the exact casing, hyphens, and version number.

For your first connection, copy an available model name from the model marketplace, then return to the example and replace YOUR_MODEL_NAME.

Model marketplace

Common errors

Symptom	Possible cause
401 or auth failure	Wrong key, key disabled, or Bearer missing from the request header
Insufficient balance	Wallet balance too low or key quota exhausted
Model does not exist	Model name misspelled, or the model is not available in the current group
Bad request format	Non-OpenAI-compatible format used, or Base URL is missing /v1