Use uncensored models in Codex

Codex reads custom providers only from its config file, not from environment variables.

Setup

Add these lines to ~/.codex/config.toml, set SILK_API_KEY in your shell, and run codex.

~/.codex/config.toml

model = "deepseek-v4.1-flash-1017-uncensored"
model_provider = "silkcompute"
model_context_window = 1048576
web_search = "disabled"

[model_providers.silkcompute]
name = "Silk Compute"
base_url = "https://api.silkcompute.com"
env_key = "SILK_API_KEY"

Keep OpenAI as your default

Save the same lines as ~/.codex/silk.config.toml instead and run codex --profile silk. Codex shows a notice that model metadata was not found. It only ships metadata for OpenAI models, and the notice is safe to ignore.