Bring your own AI to Agent host

An always-on agent calls a model many times. Someone pays for that. In Agent host, it is you, to your own provider.

Your own key, on every plan#

On Pro, Premium and Max you add your own model key. The app seals it: it is never shown again. Only your provider's model is called with it. The deployment's own keys are never used for your workspace. The plan costs the same whether your agent talks a lot or a little, and your model bill goes to your provider.

What the plans count#

The plans count agents (1, 3 or 10), how often a schedule may wake an agent, and the tool calls a day. Model use is not metered here.

A hard stop#

The tool calls a day are a hard stop, not a bill. An agent that makes too many calls stops for the day. That protects you from a loop more than any alert would.

Which setup suits you#

If you already have a model account, the setup is short: open the agent, paste the key, choose a tier and deploy. If you do not have one yet, create it with the provider you prefer, set a spending limit there, and come back. Either way the key stays yours: you can replace it at any time, and an agent without a key simply cannot answer until you add one.

A small tier is a good default for a digest or a triage agent, because those runs are short and frequent. A larger tier suits an agent that reads long documents or has to weigh several sources. You choose the tier per agent, so one workspace can hold a quick agent and a careful one side by side.

What a model provider sees#

Whatever you and the agent say, and what the tools return, is sent to the model that answers. That is your own provider under your own terms. We say this plainly because it is true of any agent.

See models and keys and usage.