Zoivu

Zoivu's homepage demo starts with one line about Aster, an elf cartographer looking for a sky island missing from her inherited atlas, with a peach braid and a weathered book. From that, the assistant writes a short character brief, draws her, and then produces a winter expedition outfit, a gouache storybook version, front, side and back reference views, a comic about the missing island and a set of sticker expressions. That chain, one character carried through many kinds of artwork, is what Zoivu is for.

You can start from words, a random prompt or a photo you have permission to use. The assistant, called Character Director, keeps the same context through the whole conversation, so you don't re-describe the character each time you ask for a new look or scene.

The feature list covers seven art directions for the same character, eight complete outfits, reference views from four angles, comic pages split into panels, storyboard keyframes for a scripted story and expression sticker sets. Higher plans allow several characters in a shared scene, up to 12 references in a request and output up to 4K.

The free tier gives 60 credits that last 14 days, which the pricing page puts at about six ordinary 1K images, and 1 GB of private storage. Starter lists at $108 a year with 500 credits a month and 1K output. Creator, Pro and Studio add more credits, parallel generations and storage, and a launch offer cuts annual plans by up to half.

The privacy policy is unusually specific. It names DeepSeek and a Langlangya gateway for text, and Agent3, ImgAPI, fal, OpenRouter and others as image routes, and says a model's display name alone doesn't identify every service that handles your data. You can ask support which supplier processed a given task. Uploads and results sit in private storage, and saved work can be exported as JSON. The service is for adults 18 and over.