Set up jcode

Merge the provider profile into ~/.jcode/config.toml. Create provider-larouter.env under ~/.config/jcode (or $XDG_CONFIG_HOME/jcode); on Windows use %USERPROFILE%\.config\jcode. Keep that key file private. Start jcode --provider-profile larouter. Reset by selecting another profile and removing the LaRouter profile and its dedicated key file.

Option 1: Use Configure

jcode separates its TOML provider profile from an environment file. This helper does not yet update that pair together; configure both below.

See manual instructions

Configuration file location

macOS: ~/.jcode/config.toml
Linux: ~/.jcode/config.toml
Windows: %USERPROFILE%\.jcode\config.toml

These are default locations. For a tool running in WSL, use its Linux path inside WSL. If you changed the configuration directory, select the tool’s actual file. Your browser can only read or write files you select.

Option 2: Configure manually

Back up your files and replace only the placeholders shown below. Codex already uses gpt-5.6-sol. Use model names without router prefixes such as cx/. Keep tool-required local provider bindings such as larouter/ where shown, and preserve unrelated settings.

Setup steps

~/.jcode/config.toml

[providers.larouter]
type = "openai-compatible"
base_url = "https://i-larouter.lapage.vn/v1"
auth = "bearer"
api_key_env = "JCODE_LAROUTER_API_KEY"
env_file = "provider-larouter.env"
default_model = "YOUR_MODEL_ID"
requires_api_key = true

~/.config/jcode/provider-larouter.env

JCODE_LAROUTER_API_KEY=YOUR_API_KEY