Understand AI tasks and assets

Generate media through visible tasks, track status and cost, and distinguish task records from reusable outputs.

Every generation is a task plus assets

Laper separates the execution record from its outputs. A generation task owns status, credit charge, estimated progress, retry behavior, and failure handling. One task can produce one or more assets. This model is shared by Script and Cinema workflows.

In a Script project, Laper Space aggregates project generation in a canvas and task list. Current production workflows include poster and casting images, trailer or scene dramatization video, and audio experiences such as director roundtables and script table reads. Available entries can evolve; use the visible task menu as the current truth.

Dispatch deliberately

Before generating, check the selected episode or entity, reference images, prompt, aspect ratio, and the credit cost shown on the action. Once dispatched, the task moves through queued, generating, completed, or failed presentation states while the backend maintains more detailed execution states.

Do not start duplicate tasks because an estimate changes or the page is refreshed. Task and asset lists reconcile from the project and can include collaborators' work.

Handle results and failures

Open the task card to inspect outputs. Assets are reusable project results; they are not automatically screenplay edits. Entity-bound workflows can route a result back to a character, prop, shot, or reel field, while project-level free creation stays in the asset library.

If a task fails, read the visible error and retry only after correcting the cause. Credits, refunds, congestion behavior, and estimated time are controlled by the current service configuration, so tutorials do not hard-code those numbers.