Start in the browser

Connect ChatGPT, Claude, then your editor

Most people chat in a browser. chatgpt.com and claude.ai cannot be pointed at Token Maxing. Put your OpenAI or Anthropic API key on this site, then chat in any browser app that has an API key and a Base URL. Cursor, Codex and Antigravity are further down.

API key tk_live_… from the Tokenmaxing email
Base URL https://api.tokenmaxing.online/v1 exactly this address — include api. and /v1

ChatGPT and Claude in the browser

Those two websites only talk to OpenAI and Anthropic. They have no Base URL field. The path below is what actually works in a browser.

OpenAI / ChatGPT

  1. Open platform.openai.com/api-keys — not chatgpt.com. Create a secret key that starts with sk-.
  2. Sign in here → Account → paste that key in the OpenAI box. That is your OpenAI account. It is not the Tokenmaxing key.
  3. Add prepaid with Stripe. When your OpenAI key pays the model, prepaid only covers Token Maxing’s 30% fee. If that OpenAI account is out of credits, the call is retried on prepaid Azure or on another connected key.
  4. In a browser chat that has API key and Base URL (TypingMind, LibreChat, Open WebUI, Chatbox): paste tk_live_ and https://api.tokenmaxing.online/v1. Pick gpt-4o or gpt-5.1.

Claude

  1. Open console.anthropic.com/settings/keys — not claude.ai. Create a key that starts with sk-ant-.
  2. Sign in here → Account → paste that key in the Anthropic box. Claude only starts on your Anthropic key.
  3. Add prepaid with Stripe. Prepaid takes the 30% fee while Anthropic bills your key. If that Anthropic account is out of credits, the call is retried on another connected key or on prepaid GPT.
  4. In the same browser chat as above: tk_live_ + the Base URL. Pick a Claude model such as claude-sonnet-5. Do not paste the Anthropic key into the chat app.

Editors — Cursor, Codex, Antigravity

Cursor

  1. Ctrl+Shift+J (Mac: ⌘+Shift+J) → Models.
  2. Turn on Use OpenAI API Key and Override OpenAI Base URL. Paste the two values above.
  3. Open a new Ask chat. Add custom model gpt-4o. Do not pick Grok or Composer.

Codex

  1. In Codex / ChatGPT Codex choose the OpenAI provider.
  2. Paste your tk_live_ key in API key.
  3. Set Base URL to https://api.tokenmaxing.online/v1.

Antigravity

  1. Open model settings and choose an OpenAI-compatible provider — not Google, not Anthropic.
  2. Paste your tk_live_ key in API key.
  3. Set Base URL to https://api.tokenmaxing.online/v1.

Gemini

Same pattern: Google AI Studio key under Account → Google. The browser or editor still uses tk_live_ plus the Base URL — never the Google key. If Gemini is not mapped on the gateway yet, the call stays on OpenAI/Azure or Claude instead.

If you want to start from a prompt

Paste this into Cursor, Codex or Antigravity. The assistant can only remind you of the two fields. It cannot flip the toggles for you.

Connect this editor to Tokenmaxing.

Use the OpenAI-compatible provider (not Anthropic, not Google):
- API key = the Tokenmaxing key that starts with tk_live_
- Base URL = https://api.tokenmaxing.online/v1

Do not select Cursor Grok, Composer, or built-in GPT-5 names. Use gpt-4o in Ask mode.
If I only have an OpenAI, Claude or Gemini key, tell me to paste that key on tokenmaxing.online → Account. The editor still uses tk_live_ + the Base URL.

Do not paste the Tokenmaxing key here

  • platform.openai.com — that site only accepts OpenAI’s own keys, not tk_live_.
  • Account on this site — the Anthropic / OpenAI / Google boxes are your provider keys, not the Tokenmaxing key.
  • You do not need a terminal or Python.

On this site first

  1. Sign in with the emailed login code (not the tk_live_ key).
  2. Add prepaid balance with Stripe. A call with $0 balance is rejected.
  3. Lost the key email? Sign in and choose Replace API key.
Python (optional)

Skip this if you chat in the browser, Cursor, Codex or Antigravity.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.tokenmaxing.online/v1",
    api_key="tk_live_paste_the_key_from_your_email",
)

ExecuteAI ApS · CVR 45798798 · info@tokenmaxing.online