Visual Novel Sprite Maker: One Character, Every Expression
Describe one character and get a standing sprite plus its expression set — same face, same outfit, same frame, ready to drop in.
A sprite is one frame of a world.
Neta Studio keeps the cast as world materials — the face, the outfit, the way they talk — so the next expression, the next background and the scene they stand in all read from the same record.
- A cast, not a folder — The character stays in the world with their look and their history, so an expression you add next week still matches.
- Scenes that answer back — The same world runs as a playable visual novel, where the character replies in their own voice instead of reading a script.
- Frames you keep — Sprites, backgrounds and the finished scene live together — export the PNGs, or keep building inside Studio.
Questions, answered
Can it make a full-body sprite, not just a portrait?
Yes. Ask for a full-body standing sprite and it comes back head to shoes on a plain background. The grid above is cut to VN framing so the faces stay legible; click any tile and the full-body master opens in the viewer.
How do the expressions stay on the same face?
The set grows from one reference frame: the first sprite is generated, then each expression is generated from it with the instruction to change only the face. In a world where the character is already saved, their cover art is the reference.
What do the files look like when they arrive?
Separate images, all the same size and ratio (about 900×1200, 3:4), on a flat light background so the edges stay clean. That is what a visual novel engine wants: one image per state, swapped by name.
Do I get transparency?
Not by default — frames come on a flat light background, which is why the edge stays clean. Ask for a cutout in the same session, or run the frames through a background remover, if your build needs alpha PNGs.
Can I use an existing character or my own OC?
Both. Name the character and the world they come from, or attach a sheet or a screenshot. Attaching a reference is the difference between "anime woman in a red dress" and your character.
Will this work in Ren'Py, Unity or Godot?
Any engine that takes a PNG. The frames are plain images named by expression, so the Ren'Py side is one image statement per state.
How many expressions can I ask for at once?
Five is the set that holds together best — neutral, smiling, blushing, furious, crying. Ask for more in the same session; each one is generated from the same reference, so the face does not drift.
What is this not?
It is not Live2D and it is not a rigged model. You get still frames, one per expression, plus new poses generated from the same character — not a blinking, breathing puppet.
Make the set before you draw the scene.
One character, five faces, one pass. The first sprite is the reference for everything after it.