← Back to image discovery

VERIFIED / Fashion & Beauty

Scroll Swordswoman by the Window

A blue-and-white swordswoman in a quiet old study, shaped by window light, dust, and tactile fabric.

PROMPT WORKSHOP

How to build the Scroll Swordswoman by the Window prompt

Do not begin with one dense block. Confirm the visual goal, then add subject, composition, light, and constraints one layer at a time.

Start with one sentence

Base prompt

Scroll Swordswoman by the Window. A blue-and-white swordswoman in a quiet old study, shaped by window light, dust, and tactile fabric. 2:3 composition.
01

Iteration

Define the result

What changes: Theme, outcome, and aspect ratio only.

Why: Check whether the model understands the core image.

Scroll Swordswoman by the Window. A blue-and-white swordswoman in a quiet old study, shaped by window light, dust, and tactile fabric. 2:3 composition.
02

Iteration

Lock the subject

What changes: Add fixed identity and scene constraints.

Why: Stable anchors reduce multi-image drift.

Scroll Swordswoman by the Window. A blue-and-white swordswoman in a quiet old study, shaped by window light, dust, and tactile fabric. 2:3 composition. Keep identity, wardrobe, props, and setting unchanged across three views.
03

Iteration

Control composition and light

What changes: Add camera, composition, and lighting constraints.

Why: Camera changes create variety while light unifies the set.

Scroll Swordswoman by the Window. A blue-and-white swordswoman in a quiet old study, shaped by window light, dust, and tactile fabric. 2:3 composition. Specify three camera distances while preserving one motivated lighting setup.
04

Iteration

Add texture and constraints

What changes: Add production detail and exclude: minor or childlike appearance, teenage appearance, nudity, transparent clothing, explicit sexual content, celebrity likeness, plastic skin, extra fingers, malformed hands, duplicate limbs, phone, screen, corner logo, floating watermark, repeated branding, unreadable text

Why: The final pass addresses common model failures.

Three cinematic portraits of the same clearly adult Chinese swordswoman in her late twenties inside an old wooden study. Mature elegant face, curvy feminine figure, full bust completely covered by layered white inner robes and a deep-blue embroidered outer robe. She holds an antique sword manual beside a carved window and a sheathed sword rests nearby. Create a window-lit cover, a seated reading detail, and a wider sword-table composition. Soft morning beams, floating dust, realistic silk and wood. Integrate exact "DUVIRO.COM" once per image as a tiny woven label on the blue sash or as restrained gold lettering on the sword scabbard; no overlay.

WHY IT WORKS

Prompt anatomy

Subject

Fixed identity, wardrobe, props, and setting across three views.

Reduces identity and object drift.
Composition

A 2:3 frame with three explicit camera distances.

Creates a coherent cover, detail, and wide sequence.
Lighting

One motivated light setup shared across all views.

Unifies material, depth, and mood.

TROUBLESHOOTING

How to fix weak results

The image looks like a generic AI template

Add one concrete camera behavior, material imperfection, and environmental detail.

Each view shows a different identity

Keep identity, wardrobe, setting, and key light fixed; change only one camera angle.

The brand looks like an overlay watermark

Integrate DUVIRO.COM once as a physical label, plaque, engraving, or debossed maker mark.

FAQ

Common questions about this prompt

How do I start a Scroll Swordswoman by the Window prompt?

Start with “Scroll Swordswoman by the Window. A blue-and-white swordswoman in a quiet old study, shaped by window light, dust, and tactile fabric. 2:3 composition.” and add identity, composition, light, and material constraints in separate passes.

Which image model should I use?

Use a current image model that handles detailed composition and identity consistency reliably.

Why do the three images look inconsistent?

Lock the face, hair, wardrobe, setting, props, and key light before changing one camera angle at a time.

Can I replace the subject or location?

Yes. Replace one variable while keeping camera and lighting stable, then evaluate before changing another variable.