From a scene to a narrated short film

Open Workflows and choose “Narrated short film”. Edit the scene description and narration independently. Use “Frame extraction and cropping” to work with an existing video.

Video and audio inputs come from your private asset library. Uploaded, generated and processed assets keep their source information.

Text to speech outputs audio. Frame extraction outputs an image; crop and resize creates a new image while keeping the original.

Combine video and audio takes video, optional audio and subtitle text. Connected audio replaces the original sound. You can extend the final frame for longer speech and explicitly enable subtitles. Lip sync is not guaranteed.

Generation nodes use enabled compatible models configured by the owner. The server validates duration, aspect ratio and credit prices. Frame extraction, cropping and composition do not spend generation credits but have storage and processing limits.

Save and confirm the step-by-step and total quote. Each admitted step is charged; unsubmitted steps are not. Failures refund the original credit source without extending gift expiry. Explicit retries retain successful results; paid requests are not automatically repeated.

Episode review shows adopted video and audio, missing assets, earlier settings and duration issues. Unknown durations remain explicitly unmeasured.

Select shots to regenerate and confirm the price. Submit speech separately in the shot editor. Adopt new candidates explicitly; existing results are not automatically replaced.