AI image generator

AI Image Generator: One Sentence, One Frame

Say who is in the picture, where they are and what the light is doing — and get the face, the light and the crop you asked for, not a lucky draw.

Boa Hancock walking along the palace terrace at sunset, seen from behind

The deck of the Thousand Sunny at night with Luffy asleep and Zoro sitting by the mast

A harbour market at golden hour with lanterns, crates and a ship at the dock

Nami steering the Thousand Sunny through a squall

Boa Hancock in a crimson dress on the Amazon Lily palace terrace at sunset — **Boa Hancock — terrace at dusk** — One sentence: who, where, what light. A finished 3:4 illustration, not a rough.

Boa Hancock at the same terrace after dark, lit by a paper lantern — **Boa Hancock — the second ask** — Same face, same dress, lanterns instead of sun: the character material carried the look.

Luffy on the lion figurehead of the Thousand Sunny in a storm — **Monkey D. Luffy — the bow** — A different character in the same world still reads as one world: one light, one palette.

Zhongli on the stone steps of Liyue Harbour at dawn — **Zhongli — Liyue steps** — Drawn from Zhongli's own character reference: same coat, same gold trim, a harbour morning around him.

Raiden Shogun standing under a maple in the Tenshukaku garden at dusk — **Raiden Shogun — Tenshukaku** — Violet dusk, katana down. The pose and the place come from the sentence, not a preset.

Nahida holding a sprout in a Sumeru greenhouse at noon — **Nahida — greenhouse** — Green noon light and a sprout in both hands; the five-colour palette was fixed before the first image.

Kyouko Hori in the school kitchen after class, apron over her uniform — **Kyouko Hori — the kitchen** — A quiet, ordinary frame. One bulb, cut-paper fabrics, no drama required.

Izumi Miyamura on the school rooftop at dusk with earphones around his neck — **Izumi Miyamura — rooftop** — A five-colour palette holds even when the scene is a school roof at dusk.

Nico Robin pulling a book from a shelf in the ship's archive at night — **Nico Robin — the archive** — Warm lamp, cold shadow: the character stays readable at grid size.

The image is made inside a world.

Neta Studio is an AI world builder, not a prompt box. The character, the place and the look are materials you keep: the face is stored on the character, the palette in the world's visual style, so the next image starts from what this one decided.

A face that holds — Ask for a second frame of the same person and the character's own art comes back, not a new guess at the face.
The frame you ask for — 16:9 for the whole terrace, 3:4 for the portrait, 1:1 for the icon. Same scene, three crops, one direction.
A reference when the face is real — Attach a photo and the model keeps the features it can see, instead of inventing a stranger with the same hair colour.

Before you type the sentence.

How is this different from a one-box image generator?

A one-box generator starts from nothing every time. Here the character, the place and the visual style already exist in a world, so the face and the light survive from one image to the next.

Can I get the same character in several images?

Yes — that is why the character is built first. The face, hair and clothes live on the character material; each new image reads them instead of guessing again from your sentence.

Which frames can I get?

16:9 for a wide establishing shot, 3:4 for a portrait, 1:1 for an icon or a profile picture — plus 4:5, 2:3 and 9:16 when a platform needs them.

Can I use a photo of a real person?

Attach it and say what matters: face shape, hair, glasses, a scar. The model keeps what it can see. Use a photo you have the right to use.

Why did the face come back wrong?

Usually because the character was only described in the prompt, or the scene was so wide the face had twelve pixels. Define the character in the world first, then ask for the frame.

Can I put text in the image?

Short labels sometimes land; longer lines rarely do. Generate the picture clean and set the type in the page around it.

Where do the images go?

Each one is saved as a work in the world at full size, together with the sentence that made it. You can reopen it, re-run it, or put it on the board next to the others.

Do I own what I generate?

You own what you bring — the world, the characters, the direction. The model's own terms decide the rest; check them before you sell an image.

One sentence. One finished image.

Name the character, the place and the light. The frame comes back at the size you asked for.

Generate My Image