How to set up Codex CLI with the Tokenator API

What you end up with

Once configured, Codex CLI stops talking to its default provider and sends every request to Tokenator. Billing comes out of your key's token limit, and the list of available models is defined by the key rather than by the tool.

Requirements and supported operating systems

• Windows 10/11, macOS 12+, or a modern 64-bit Linux.
• Outbound HTTPS access to the Tokenator domain.
• A Tokenator API key that looks like sk-your-tokenator-key.

Installing via npm requires Node.js LTS. Codex CLI also ships standalone binaries, in which case Node is not required.

Installing Codex CLI

npm install -g @openai/codex

brew install --cask codex

Run codex once after installing so the ~/.codex directory is created.

Setting up Tokenator

1. Register in your account and buy a token bundle.
2. The key appears in the account right after the payment is confirmed — copy it whole, including the prefix.
3. The examples below use the placeholder sk-your-tokenator-key. Never publish a real key in repositories or screenshots.

Configuration on Windows

model = "gpt-5.5"
model_provider = "tokenator"

[model_providers.tokenator]
name = "Tokenator"
base_url = "https://api.tokenator.top/v1"
env_key = "TOKENATOR_API_KEY"
wire_api = "responses"

$env:TOKENATOR_API_KEY = "sk-your-tokenator-key"
codex

setx TOKENATOR_API_KEY "sk-your-tokenator-key"

The provider id must not collide with the reserved ones (openai, ollama, lmstudio), hence tokenator in the example. The key never goes into the file: env_key names the environment variable to read it from.

Configuration on macOS and Linux

model = "gpt-5.5"
model_provider = "tokenator"

[model_providers.tokenator]
name = "Tokenator"
base_url = "https://api.tokenator.top/v1"
env_key = "TOKENATOR_API_KEY"
wire_api = "responses"

export TOKENATOR_API_KEY="sk-your-tokenator-key"
codex

The provider id must not collide with the reserved ones (openai, ollama, lmstudio), hence tokenator in the example. The key never goes into the file: env_key names the environment variable to read it from.

Your first request

Before starting the tool it helps to confirm that the key and the model both work. A single curl request answers both questions.

curl https://api.tokenator.top/v1/chat/completions \
  -H "Authorization: Bearer sk-your-tokenator-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.5",
    "messages": [{"role": "user", "content": "ping"}]
  }'

cd /path/to/project
codex

A successful response means the key is valid, the model is allowed for it and the upstream is reachable. Usage appears in your account right after the request.

Switching models

Change the model with the model field in config.toml, or per run with codex -m <model-id>.

Examples of model IDs available right now:

• claude-haiku-4-5
• claude-opus-4-6
• claude-opus-4-7
• claude-sonnet-4-6
• gpt-5.5
• deepseek-v4-flash

Current IDs and specs live in the AI model catalog, which also shows which models are available right now. The model comparison helps choose between close options.

Configuring reasoning

Codex CLI passes the reasoning effort in the request itself — in the config that is model_reasoning_effort. The value is forwarded upstream as-is; models that do not support it ignore it without erroring.

Reasoning is billed as output tokens, so a high level noticeably increases usage.

Common errors and how to fix them

FAQ

Read next