AI on Cloud
When your workspace's plan includes AI and the server offers it, the AI assistant can use models through Cloud instead of a key of your own. In Settings → Models, add a model, choose OpenPencil Cloud as the provider, and pick the workspace whose plan pays for it; the models and their prices come from the server. OpenPencil Cloud is listed while you are signed in to the server Home shows.
Each request uses the credits of the workspace owner's plan for the month, which every workspace they own shares, so requests in one can use up the credits of the others. When they run out, the assistant says so and, if the server bills, offers Manage plan. Self-hosted servers run their own AI service with their own models and limits (self-hosting guide).