Most people discover styles by accident — a stray word like “illustration” tips a render from photo into painting, and they reverse-engineer from there. There is a faster way. Styles respond to anchors: short, high-signal phrases that name a medium, an era, or a production pipeline. Learn a handful and you can art-direct on purpose.
To demonstrate, we rendered the identical subject — “a red fox sitting in a forest clearing, morning light” — through four anchors. Same model, same seed discipline, only the style words changed.
Photoreal
…natural photography, 85mm lens, shallow depth of field, golden hour, fine fur detail
The anchor here is “natural photography” plus lens vocabulary. Naming a focal length does more than any number of “ultra realistic” adjectives — it tells the model which physical camera system to emulate, and realism follows. Avoid the word “render” entirely; it drags results toward CGI.
Anime
…anime key visual, cel shading, bold linework, saturated colors, dramatic sky
“Key visual” is the power phrase — it invokes the polished promotional art style of anime production, rather than generic manga sketches. “Cel shading” locks the flat-color look; “bold linework” keeps edges graphic instead of painterly.
Watercolor
…loose watercolor painting, wet-on-wet washes, visible paper texture, soft edges, minimal detail
Watercolor lives or dies on restraint. “Wet-on-wet” and “visible paper texture” are the anchors that produce genuine medium behavior — pigment blooming, white paper breathing through. Asking for “minimal detail” is counterintuitive but crucial: overworked watercolors read as digital paintings.
Stylized 3D
…stylized 3D render, subsurface scattering, soft studio lighting, Pixar-style character design
Here “stylized 3D” matters more than “3D” alone, which can drift photoreal. “Subsurface scattering” gives materials that soft, lit-from-within quality of feature animation, and naming a studio style sets the whole character-design vocabulary at once.
Why anchors beat adjectives
Adjectives describe quality (“beautiful”, “detailed”, “stunning”) and models mostly shrug at them — everything in the training set was somebody’s idea of beautiful. Anchors describe provenance: what medium this image was made in, by what kind of artist, for what kind of audience. That is information a model can actually act on.
The four presets above ship as one-click styles in the Supervivid image editor, but the deeper point is transferable: whatever look you are chasing, find the words a practitioner of that craft would use, and put those in the prompt.