Codex Integration

Configure as a model provider in the Codex CLI.

Recommended: use env_key — Codex reads the environment variable at startup and sends it as a Bearer Token. Alternatively, use experimental_bearer_token to embed the key directly. Use one or the other, not both.

Recommended: config.toml with env_key

model = "gpt-5.5"
model_provider = "neotoken"

[model_providers.neotoken]
name = "Neotoken"
base_url = "https://your-router-domain.com/v1"
env_key = "NEOTOKEN_API_KEY"
wire_api = "responses"
supports_websockets = false

Alternative: config.toml with Bearer Token

model = "gpt-5.5"
model_provider = "neotoken"

[model_providers.neotoken]
name = "Neotoken"
base_url = "https://your-router-domain.com/v1"
experimental_bearer_token = "YOUR_API_KEY"
wire_api = "responses"
supports_websockets = false

Config file location

macOS

Config file is usually at ~/.codex/config.toml .

Open in terminal:

mkdir -p ~/.codex && open -e ~/.codex/config.toml

Set for current shell session:

export NEOTOKEN_API_KEY=YOUR_API_KEY

For a permanent setting, add it to ~/.zshrc, then reopen the terminal or run source ~/.zshrc.

Windows

Config file is usually at %USERPROFILE%\.codex\config.toml .

Open with PowerShell:

mkdir $env:USERPROFILE\.codex -Force; notepad $env:USERPROFILE\.codex\config.toml

Set for current PowerShell session:

$env:NEOTOKEN_API_KEY = "YOUR_API_KEY"

For a permanent setting: [Environment]::SetEnvironmentVariable("NEOTOKEN_API_KEY", "YOUR_API_KEY", "User") — reopen the terminal for the change to take effect.

Which should I use?

env_key is more secure — the key is never written to config.toml. Recommended for everyday use.
experimental_bearer_token is simpler to configure but writes the key to config.toml. Only recommended for local debugging or controlled environments.
After changing the environment variable or config.toml, restart Codex for the new settings to take effect.

Image / Video Tasks Claude Code