Skip to main content
An AI Provider is the canonical Clawdi object that records how an Agent reaches a model service. A Cloud Agent does not have a different kind of AI Provider; its AI Providers section binds that shared concept to the remote runtime.

Choose an access mode

The dashboard marks providers that are unavailable for the current runtime and limits the selectable models to the relevant catalog. Do not assume universal compatibility between an authentication method, API protocol, model, and every Agent software choice.

Configure access during deployment

In the deployment wizard:
  1. Choose the Agent software first.
  2. Choose Clawdi AI, a usable saved provider, or Configure inside agent.
  3. Select the primary model when the chosen provider requires one.
  4. Review the selection before deploying.
The CLI offers the same conceptual choices through the interactive wizard. For automation, use clawdi deploy --help from the installed CLI and pass saved provider identifiers exactly as shown. Provider secrets are not accepted as deployment flags.

Change a running Agent

Open the Cloud Agent, choose AI Providers, select the new provider and primary model, then choose Save changes. The Agent can enter Updating while Clawdi applies the binding. Wait for that change before submitting another provider or lifecycle action. Account-level AI Providers is where you add, validate, edit, or remove saved provider connections. The Agent’s AI Providers section only selects which connection and model that remote runtime should use.
The local AI Provider catalog used by a Connected Agent and the Cloud Agent controls in the dashboard are different surfaces. Changing one does not automatically change the other.
If a saved provider becomes unavailable or is removed, there is no automatic fallback to Clawdi AI. Fix or re-add the account provider, choose another usable provider, choose Clawdi AI, or configure access inside the Agent. For more recovery steps, see Cloud Agent troubleshooting.
Last modified on August 2, 2026