Choosing a default model

Each agent has a default model — the model it uses when a request doesn’t ask for a specific one. You can change it at any time, and the change takes effect without reinstalling integrations or restarting from scratch.

Where model configuration lives

Model choice is core agent configuration, not a separate integration you install. Every agent is set up with access to the supported models out of the box, so there’s no “connect a model provider” step — you pick which model should be the default.

You manage this from the agent’s settings in the dashboard, where you’ll see the available models grouped by provider and can select the one to use by default.

Setting the default from the dashboard

1. Open the agent in the dashboard.
2. Go to its model configuration.
3. Pick the model you want as the default and save.

The agent starts using the new default going forward. Because model routing is centralized, you don’t need to change any channels, integrations, or workspace files to switch models.

Using a different model per request

The default is just the fallback. An agent can use any available model on a per-request basis, so you’re free to route routine work to a faster, cheaper model and reserve a stronger model for harder tasks — all under the same agent.

Restricting which models are available

Account administrators can limit which models the agents in their account are allowed to use. When a restriction is in place, the model picker shows only the approved models, and requests for models outside that set are rejected. If you expected a model to be available and it isn’t, check with your account administrator.

What’s next

• Bring your own key — point a provider at your own account.
• How model usage is billed — what model usage costs against your credit pool.