GPT Image 2 · 1930s Cartoon Key Visual

GPT Image 2 1930s Cartoon Key Visual Generator

A sepia filter turns a modern picture brown. Studio holds the 1930s drawing rules — noodle limbs, pie-cut eyes, a flat cream sky — in one wide frame, drawn by GPT Image 2.

1930s rubber-hose cartoon of Miyamura in an empty classroom at night, one desk lamp lit and a full moon in the tall window

1930s rubber-hose cartoon of Ishikawa laughing by the bicycle shed at sunset, cherry trees and the school gate behind him

1930s rubber-hose cartoon of Raiden Shogun on the black Inazuma cliffs, a naginata held low and one violet lightning bolt cracking the sky

1930s rubber-hose cartoon of Homelander hovering over a ticker-tape parade, arms wide and red cape snapping, a cartoon New York street below

1930s rubber-hose cartoon of Venti the bard flying with his lyre above Mondstadt, the cathedral and a red sunburst behind him — **Venti · Genshin Impact** — Drawn by this page's own button inside a world that already held the Genshin cast: the lyre, the cathedral and the Mondstadt rooftops came from the world.

1930s rubber-hose cartoon of Gojo in a black uniform under a cracked Shibuya screen tower, cartoon cars and shopping bags in the rain — **Satoru Gojo · Jujutsu Kaisen** — Shibuya as a cartoon street: cracked screen tower, cars with eyes, one black uniform against cream sky.

1930s rubber-hose cartoon of Sukuna with four arms on a heap of rubble, a red sun disc and skulls behind him — **Ryomen Sukuna · Jujutsu Kaisen** — Four arms and four gloves on one body, a flat red sun doing the whole backdrop.

1930s rubber-hose cartoon of Yuji throwing a punch with a stretched arm, a shuttered Tokyo street and a vending machine behind him — **Yuji Itadori · Jujutsu Kaisen** — The rubber-hose stretch used as action: the arm crosses the whole frame in one curve.

1930s rubber-hose cartoon of Homelander in a blue suit and red cape in front of the Vought tower at dusk, a cheering crowd at his feet — **Homelander · The Boys** — A photoreal suit turned into flat cel: eagle on the shoulder, cape as one red shape.

1930s rubber-hose cartoon of Starlight waving on a vaudeville stage, red curtains and a row of blank marquee bulbs behind her — **Starlight · The Boys** — The heroine recast as a vaudeville act — footlights, curtains, blank bulbs for lettering.

1930s rubber-hose cartoon of Hori sitting on a rooftop ledge with a paper lantern, a water tower and festival lanterns behind her — **Hori Kyouko · Horimiya** — A slice-of-life character in the same ink: one warm lantern against a teal evening.

1930s rubber-hose cartoon of Nahida floating among giant mushrooms and vines in the Sumeru rainforest, a pale tree glowing behind her — **Nahida · Genshin Impact** — Sumeru as a stage flat: giant mushrooms, a pale trunk, fireflies as flat yellow dots.

1930s rubber-hose cartoon of Zhongli at the rail of Liyue Harbour at dusk, pagodas and junks behind him and lanterns rising — **Zhongli · Genshin Impact** — Liyue at dusk, the harbour carrying as much of the frame as the man at the rail.

A drawing rule, not a filter

Neta Studio is a world builder: the cast, the places and the art direction live in one world, and every frame is drawn from them. The 1930s rules — wobbly ink, noodle limbs, flat cream, one hard accent — are written once, so the next character arrives in the same cartoon language instead of a fresh guess.

The cast stays on-model — Attach a character sheet, or let the world's own cover stand in, and hair, face and costume survive the restyle.
A palette you can repeat — Cream paper, soot black, umber, faded barn red, dusty teal — the same five notes on every frame of a set.
Wide frames, quiet top — 16:10 with the upper band left pale, so a title, a logo or a storefront capsule has somewhere to sit.

Questions before you start

Is this a 1930s photo filter?

No. A filter tints a photo you already have. This draws a new frame: the body is rebuilt with noodle limbs and white gloves, the eyes become pie cuts, and the background is drawn flat.

What makes it read as 1930s and not just retro?

Three things carry it: continuous rubber-hose limbs with no elbow or knee, four-finger gloves and overshoe feet, and pie-cut eyes. Then the print — uneven ink contour, flat cel fills, aged cream paper, lithograph grain.

Which model runs it?

GPT Image 2, one frame per run at a wide 16:10 size. The prompt names the size because the model defaults to square.

Will my character still look like themselves?

With a reference, yes: attach a sheet or a portrait and the prompt tells the model to keep the face, hair and colours. Without one it falls back to the character's cover in the world, or invents a face that fits the style.

Can it restyle a live-action character?

Yes. Homelander and Starlight above started as photoreal references; the restyle keeps the costume cues and drops the realism.

How many tries does one frame take?

Usually one. The frame that comes back can crowd the title band, so the second pass is a small edit — lift the figure, deepen the shadow side — not a new prompt.

Why is there no text on the posters?

Lettering comes back garbled more often than not. The prompt asks for a pale, quiet band instead, and you set the title yourself in any editor.

Can I build a whole set for a game or a pitch?

Yes. Change the subject line and keep the closing rules: the palette and ink weight hold across characters, which is what makes nine frames read as one project.

Draw the first frame

One prompt, one wide key visual, and the same rules ready for the next character.

Draw My Key Visual