Characters API

What the characters endpoints are for

A character is a reusable presenter: a face and a voice you can put in video after video instead of getting a different stranger each render. This is what makes a series of videos look like they come from the same channel.

The creation flow is deliberately two-stage. POST /characters/variations generates portrait candidates and POST /characters/voice-previews auditions voices reading a sample line, so you choose from real options rather than describing what you want and hoping. Once you have a face and a voice you like, POST /characters saves the pairing to the workspace.

Generating portrait candidates spends credits, because it renders images. Listing, updating and deleting saved characters does not.

Characters API questions

How do I keep the same presenter across multiple videos?

Save a character once with POST /characters, then reference it on each render. Characters persist in the workspace, so every video that uses one gets the same face and voice.

Can I hear a voice before committing to it?

Yes. POST /characters/voice-previews returns candidate voices reading a sample line, so you audition them against your own copy rather than picking from names.

Does generating a character cost credits?

Generating portrait candidates does, since it renders images. Saving, listing, updating and deleting characters does not.

Can I use my own photo instead of a generated portrait?

Yes. POST /characters accepts uploaded assets as well as generated ones, so a character can be built from an image you supply through the uploads endpoints.