The Inference API is designed to call ready-to-use models through an OpenAI-compatible interface rather than launching your own instance. If the option is missing or marked closed in Token market (/inference), it is not open for your account. When enabled, this page lists models, availability, and recent requests. Do not assume that a model shown in an example is currently available.
What is the Inference API?
From the EverXCloud help center.
