Open Environment, then Models & connections. Find provider credentials below the engine connections. This section configures API models, separately from subscription-based engines.

Choose Platform key to use deployment credentials, or My key for your own credentials. Switching sources preserves the other saved key, but does not combine them.

Enter your provider key under My key. Change the optional base URL only when needed for your endpoint. Save before testing the connection.

Select Test connection and read the provider result. If it fails, check the selected source, key, and endpoint, then save the correction before retesting.

Expand Model catalogs & visibility. Working credentials do not make every model visible. Enable the models you want to appear in the chat picker.

For OpenRouter or SiliconFlow, refresh the model whitelist, search it, and enable your choices. The provider’s current catalog determines what is available.

Return to a new chat and select the enabled model under API models. In your own workspace, send a simple question to verify the setup.