Skip to content
Joyland

AI Chat With Images

Some pictures identify a character; others belong to a particular moment in the story. Decide which job an image needs to do before choosing a visual feature.

Three different kinds of picture

A portrait introduces the cast. A gallery contains pictures associated with a character. An image message is part of a conversation. Access to one doesn't necessarily include the other two, so check the allowance attached to the action.

You can view the existing portraits in our character collection without requesting a new picture. For AI image messages or gallery unlocks, consult the credit and membership comparison rather than assuming that free text access covers every visual option.

Describe what the story needs to show

Begin with the subject and location, then add one detail the dialogue can use. For a gallery mystery, the position of a label or an empty frame may be more important than elaborate lighting effects. Keep established appearance details consistent with the character.

A gallery attendant beside an empty picture frame, holding a visitor's photograph that shows the same wall before the opening.

Our scene-note tool can organize a written idea. It produces text, not an image. There is no connected picture-generation request in the browser drafting tools.

Use material you have a right to share

Choose your own images or material you are permitted to use. A downloadable portrait isn't automatically free of usage restrictions. Don't use a private person's likeness for an intrusive scenario or upload someone else's private photos.

If the picture is meant to explain a clue, describe that clue in the writing as well. The scene should remain understandable to somebody who can't see the image.

A few things to know

Does opening a portrait spend generation credits?

Viewing an existing picture and requesting a newly rendered one are different actions. Check the service's displayed cost for the specific request.

Can the character editor generate a portrait?

No. The editor here saves a text card. It doesn't render images or send prompts to an image model.