How to Configure Codex + a Model

‍

Codex (the command-line coding model) is configured under Admin Center → Model Integration → Codex. Each "channel" maps to a ~/.codex/<channel>.config.toml file and is selected at session start via codex --profile <channel>.

1. Open Model Integration → Codex

• Top-right avatar menu → Admin Center → the Model Integration tab → pick Codex at the top right. The left side lists imported channels; the right side is the add / edit form.

2. Add a Codex Channel

• Click New, then fill in the fields below (it only saves — nothing is invoked yet):

Field	Meaning
① Channel (ASCII letters only)	Channel name; also the --profile id and config filename, e.g. mychannel
② Display name	Name shown in the model picker, e.g. My Codex
③ Codex model name (-m)	Passed to codex as -m, e.g. gpt-5.5
④ Secret name (env_key)	Env var the api_key is exported to; defaults to RIGHTCODE_API_KEY
⑤ Secret value	The real API key (sk-...); leave blank when editing to keep the old one
⑥ config TOML	Auto-generated; written to ~/.codex/<channel>.config.toml with base_url / api_key

• Click Save when done.

3. Use It

• Saving writes ~/.codex/<channel>.config.toml and exports the secret to your env_key.
• The channel then appears in the model picker when creating a new session / research task. In the list, the three badges enabled / config written / secret set all being green means it's ready.