Character models
A character model represents a reusable identity for supported generation workflows. It helps you create new media featuring that character. Creating a model and generating an image are separate operations. Model creation has its own task and readiness state. Wait for the model to become ready before using it in a generation that requires a trained model. Some workflows can generate media without a character model. Follow the choices offered by your integration rather than assuming every operation requires one.A Portrait Master image is an image output. Creating one does not, by itself,
create a trained character model.
Reference elements
An element is a reusable reference asset. Depending on the workflow, a reference can provide visual content, motion, or audio context. Personal elements belong to your account. Community elements are usable only when their visibility and availability permit it. A known identifier is not enough to access another account’s private reference. A workflow may support image, video, or audio elements with different limits. These limits can depend on quality, duration, source footage, and voice selection. Use the options exposed by the selected workflow; a reference accepted by one workflow is not automatically valid in another.Generated media
Media is an individual image, video, or audio output. A single generation can produce multiple media items. Keep the output you want to reuse and select it as the source of a later operation where supported.Which resource do you need?
Learn how these resources relate to tasks and results,
then prepare your input media.