Skip to content
Documentation
Browse documentation

Configure an AI provider

Set up provider credentials, an endpoint, and models, then try a conversation.

Updated
On this page

Prepare connection details

Get an API key, endpoint, and available model IDs from your provider. Treat the key as a credential and keep it out of public documents and screenshots. A display name may differ from a model’s actual ID.

Add a provider

  1. Open Settings from the AI chat toolbar, select Providers, and choose Add provider.
  2. Select the provider type and enter a name, API key, and the required endpoint.
  3. Configure available models and choose a default. Select a protocol when required by the provider’s interface.
  4. Save, return to chat, choose a model under My Providers, and send a short message.

Troubleshoot a connection

For authentication errors, check the key. For unknown-model errors, check the model ID. For connection failures, check the endpoint, network, and proxy. Compatible interfaces can still differ in protocols and model capabilities; use the provider’s own configuration details.

The built-in provider is managed through your account, separately from your custom configurations. See Built-in AI.

Report a documentation issue

Something unclear or different in the app? Let us know.