Prompting for believable scenes

Describe each fictional adult separately, then define the shared setting, lighting and interaction. This reduces blended identities and helps the model preserve distinct faces, clothing and body types.

Avoid stereotypes and use relationship or mood language that emphasizes mutuality rather than reducing characters to a category.

Choosing the right workflow

Image generators provide more control for two-character compositions; video generation adds temporal complexity and identity drift. Begin with a strong still composition before attempting animation.

Use our image generator guide and reviewed platform pages to compare reference and variation controls.

Consent remains central

Only use fictional adults or source material with clear permission from every depicted adult. Do not infer consent from a public photograph.

Use only images of consenting adults aged 18 or older. Never create intimate media of a real person without explicit permission, and review the law and platform rules that apply where you live.

Recommended next step

Compare the image, video and GIF workflows, then read a provider review before registering.

Frequently asked questions

Why do two-character images merge faces?

Multi-subject scenes are harder for models. Separate descriptions, spatial language and reference controls can help.

Is a public image acceptable as a reference?

Public availability is not consent for sexual transformation. Obtain explicit permission first.

Editorial note

This guide is educational and does not endorse non-consensual deepfakes. Features, prices and policies change; verify current information on the linked provider pages.