Candy AI anime: styling illustrated companions
Anime here is not a different product. It is a rendering style laid over a companion you already have, which means the interesting work is in the wording. Four short clauses do it. Longer prompts do it worse.
Generate an illustrated lookA styling choice, not a separate app
The image generator sits inside the chat, and the character behind it does not change when you move to illustrated output. Her personality, her selected voice and everything the persistent memory layers have picked up stay exactly where they were. You are changing the drawing, not the person.
So a companion built on the girlfriend side, or a male one from the boyfriend page, can be rendered illustrated without losing a single exchange of history.
The four clauses
Illustrated styles are less forgiving than photoreal ones. A vague description shows up as a face that drifts, so keep the prompt short and in this order.
1.Who she is
A name plus the two or three traits that define face and hair. Write this once, then paste it unchanged into every later prompt. It is the anchor.
2.Where she is
One place, one time of day. A rooftop at dusk gives the generator more to work with than "somewhere pretty" ever will.
3.How close the frame is
Portrait, waist-up, full body. Say it. Left unsaid it changes between renders, and a set stops looking like a set.
4.One adjective of mood
Exactly one. Stacking contradictory moods is the single most common reason an illustrated render comes back muddy.
Keeping her the same
Treat the first render you like as the reference. Copy the exact clause that produced the face and reuse it word for word; change the setting and the framing and nothing else. People who improvise on every prompt end up with a gallery of near-misses and blame the generator.
Prove a new idea on one image. Scale it after, never before.
The studio roster is a good place to see what a consistent look reads like, and what high-quality generation needs from an account is on the pricing page. The home page has the shorter version of all of it.
Questions
Is anime a separate app or a setting?
A setting. The companion behind the picture keeps her personality, her voice and her memory of your conversations, so you can take a character you already talk to and ask for illustrated images instead of photoreal ones.
Why do two renders look like different people?
Because the prompt moved. The generator rebuilds the image from the text you give it each time, so paraphrasing the hair and eye detail is enough to shift the face. Keep a fixed character clause and vary only the setting.
Does an illustrated companion chat differently?
No. Styling changes what the image generator outputs, not how she writes. Adaptive responses, persistent memory and the voice you selected all behave the same way whether the character is drawn or photoreal.
What do I need to generate images?
High-quality image generation sits on the premium side of the account, alongside unlimited messaging and voice notes. A free account is enough to test chat but restricts advanced features.