Guides · Codex CLI

Codex CLI on Sideren.

Codex talks the OpenAI Responses protocol. One provider block in config.toml points it at the gateway; the key travels through an environment variable.

One command

npx sideren setup --tool codex --key sdn_your_key

The CLI writes the provider block below into ~/.codex/config.toml (with a backup) and prints the export line for your key. See CLI.

By hand

Add a provider to ~/.codex/config.toml. Two details matter: the base URL includes /v1, and the wire protocol is responses.

[model_providers.sideren]
name = "Sideren"
base_url = "https://api.sideren.io/v1"
env_key = "SIDEREN_API_KEY"
wire_api = "responses"

model_provider = "sideren"
model = "sdn-gpt-oss-120b"

~/.codex/config.toml

Then export the key Codex will read:

export SIDEREN_API_KEY=sdn_your_key

model accepts any catalog id — a coding flagship like sdn-glm-5 is a strong daily driver for heavy sessions.

Troubleshooting

Empty responses? Reasoning models spend hidden reasoning tokens from max_output_tokens — a very low budget can be consumed before any text is emitted. Raise it (4,000+ is a safe floor) or pick a non-reasoning model.
401 unauthorized? Codex reads the key from the variable named in env_key — make sure SIDEREN_API_KEY is exported in the shell that launches it.
404s? Unlike Claude Code, the Codex base URL must end in /v1.
Protocol errors? Keep wire_api = "responses" — the gateway serves the Responses dialect on /v1/responses for exactly this client.

Note

Switching models mid-project is a one-line change to model — the provider block stays put.