Codex CLI

Codex CLI can use a custom model provider. Tokeness connects via the OpenAI Responses protocol — the Base URL is always:

https://n.tokeness.dev/v1

If you use CC Switch, it writes Codex's auth.json and config.toml for you. If you need to check the configuration manually, this page also provides file examples.

Connection details

Field	Value
Provider Name	Tokeness
Base URL	https://n.tokeness.dev/v1
API Key	A key created in the Tokeness console
Model	Copied from the Tokeness model marketplace
Wire API	responses

1. Prepare Tokeness

Prepare three things first:

1. A Tokeness API key.
2. A Tokeness model name.
3. Available balance.

Create a separate key for Codex:

codex-cli-local

If you later notice abnormal Codex consumption, you can disable just this key without affecting other tools.

2. Install Codex CLI

Confirm Node.js is available:

node --version
npm --version

Install:

npm install -g @openai/codex

If npm downloads are slow:

npm install -g @openai/codex --registry=https://registry.npmmirror.com

Confirm the command is available:

codex --version

3. Configure through CC Switch

1. Open CC Switch.
2. Switch to Codex on the left.
3. Click the + button in the top right.
4. If there is no Tokeness preset, choose a custom provider.
5. Set the name to Tokeness.
6. Enter your Tokeness API key.
7. Set the Base URL to https://n.tokeness.dev/v1.
8. Enter the full model name from the Tokeness model marketplace.
9. For API type or Wire API, select responses.
10. Save and enable.

After switching, close the current terminal, reopen one, and then run Codex. Codex CLI usually reads its configuration at startup, so the old terminal may still hold the previous environment.

4. Manual configuration reference

If you want to verify what CC Switch wrote, you can check the Codex configuration directory.

macOS / Linux:

ls ~/.codex

Windows PowerShell:

Get-ChildItem $env:USERPROFILE\.codex

auth.json example:

{
  "OPENAI_API_KEY": "your Tokeness API key"
}

config.toml example:

model_provider = "tokeness"
model = "the model name copied from the Tokeness model marketplace"
disable_response_storage = true

[model_providers.tokeness]
name = "Tokeness"
base_url = "https://n.tokeness.dev/v1"
wire_api = "responses"
requires_openai_auth = true

Field descriptions:

Field	Description
model_provider	The currently active provider
model	The model Codex uses by default
base_url	Tokeness endpoint address
wire_api	Codex uses the Responses protocol
requires_openai_auth	Uses OpenAI-style Bearer key authentication

If you manage Codex through CC Switch, do not manually edit these two files long-term. The next time you switch providers in CC Switch, your manual changes may be overwritten.

5. Verification

Reopen the terminal:

codex

Enter:

Please reply with a single sentence confirming Codex is now connected to Tokeness.

Then check the Tokeness usage logs to confirm:

• A new request appeared.
• The key is the Codex-dedicated key.
• The model name is correct.
• The status code is successful.

6. Multi-model configuration

You can create multiple Codex providers in CC Switch, or switch the model in the configuration.

Configuration name	Purpose
Tokeness Codex Coding	Daily code modifications
Tokeness Codex Reasoning	Complex analysis, solution design
Tokeness Codex Docs	Document cleanup, bulk rewriting

After switching, reopen the terminal.

7. Troubleshooting

Symptom	Fix
codex command not found	Reinstall, or restart the terminal
Still using the old model	Close the terminal and reopen; check model_provider
401 Unauthorized	Check the key in auth.json
404 Not Found	Check that base_url is https://n.tokeness.dev/v1
Responses-related errors	Check that wire_api = "responses"
model not found	Re-copy the model name from the Tokeness model marketplace
No Tokeness logs	Codex is not using the Tokeness configuration; check whether the provider is enabled