Codex CLI (Server)

Deploy the Codex CLI on a WSL2 or SSH server and connect to the LMU AI API: install Node 20 with volta, then write config.toml and auth.json in one command.

For configuring the Codex CLI in a WSL2 or SSH remote server environment.

Configuration steps

Step 1 — Install the Node.js environment (with volta)

In the server terminal, run:

curl https://get.volta.sh | bash

Step 2 — Install Node 20

Note

After volta installs successfully, you must open a new terminal window before running the following command:

volta install node@20

Step 3 — Install the Codex CLI

volta install @openai/codex

Step 4 — Create the config files

Run the following in the server terminal to write both ~/.codex/config.toml and ~/.codex/auth.json at once (remember to replace the key in auth.json with the sk- key generated in the console):

mkdir -p ~/.codex

cat > ~/.codex/config.toml <<'EOF'
model = "gpt-5.4"
model_reasoning_effort = "xhigh"
disable_response_storage = true
sandbox_mode = "danger-full-access"
windows_wsl_setup_acknowledged = true
approval_policy = "never"
file_opener = "vscode"
model_provider = "codex"
web_search = "cached"
suppress_unstable_features_warning = true

[history]
persistence = "save-all"

[tui]
notifications = true

[shell_environment_policy]
inherit = "all"
ignore_default_excludes = false

[sandbox_workspace_write]
network_access = true

[features]
plan_tool = true
apply_patch_freeform = true
view_image_tool = true
unified_exec = false
streamable_shell = false
rmcp_client = true
elevated_windows_sandbox = true

[profiles.auto-max]
approval_policy = "never"
sandbox_mode = "workspace-write"

[profiles.review]
approval_policy = "on-request"
sandbox_mode = "workspace-write"

[notice]
hide_gpt5_1_migration_prompt = true

[model_providers.codex]
name = "codex"
base_url = "https://api.lmuai.com"
wire_api = "responses"
requires_openai_auth = true
EOF

cat > ~/.codex/auth.json <<'EOF'
{
  "OPENAI_API_KEY": "your sk- key generated in the console"
}
EOF

The config contents are identical to the Codex CLI (Mac/Linux) guide — you can also create the two files by hand with vim / nano.

Step 5 — Launch the Codex CLI

codex --dangerously-bypass-approvals-and-sandbox

Or:

codex --yolo

Enabling the 1M context (optional)

gpt-5.4 supports a 1M-token long context, but it's off by default. Edit ~/.codex/config.toml, add two lines at the top level of the file (before any [section]), and re-enter Codex:

model_context_window = 1000000
model_auto_compact_token_limit = 900000

For field meanings, how to verify, and caveats, see Enabling the 1M context in Codex.

Last updated: 2026-08-26