Eight original, production-oriented Qwen Image 3.0 prompt recipes. Replace bracketed variables and place final on-image copy in Text (verbatim).
Use it for: portray [SUBJECT DESCRIPTION] as observant and self-possessed in a real evening neighborhood
Target language: English (en)
Localization directive: Convert every user-editable variable and non-English quoted example into natural English before rendering. Preserve deliberately multilingual copy only when the recipe explicitly requires it. Keep authorized brand and product identifiers unchanged.
Typography QA: Use correct English spelling, punctuation, capitalization, and natural phrase-based line breaks. Treat non-English Text (verbatim) content as source copy to localize unless the recipe is explicitly multilingual. Do not add filler text.
Use case: commercial-photography
Asset type: 4:5 editorial street portrait for a magazine profile
Goal: portray [SUBJECT DESCRIPTION] as observant and self-possessed in a real evening neighborhood
Scene/backdrop: narrow corner shop at blue hour, cool skylight outside, warm fluorescent spill inside, light rain on pavement, no identifiable storefront branding
Subject: one authorized adult subject in a simple charcoal coat, hands naturally occupied with a folded newspaper, direct but relaxed gaze
Camera language: full-frame documentary camera, 50 mm lens, eye-level, f/2.8, 1/160 s feeling, restrained depth of field
Lighting/mood: preserve believable mixed color temperatures; soft face exposure without beauty-light flattening; reflective and intimate
Texture: visible pores, fine hair, damp wool, subtle sensor grain, realistic highlight roll-off
Color grade: cool cyan shadows, warm practical lights, neutral skin, no orange-teal exaggeration
Constraints: one person; anatomically correct hands; plausible reflections; natural facial asymmetry; no real shop logos; authorized subject only
Avoid: plastic skin, fashion-runway posing, artificial bokeh shapes, duplicated pedestrians, HDR halos, watermark
Use it for: make a freshly served [DISH] feel immediate, honest and locally grounded rather than studio-perfect
Target language: English (en)
Localization directive: Convert every user-editable variable and non-English quoted example into natural English before rendering. Preserve deliberately multilingual copy only when the recipe explicitly requires it. Keep authorized brand and product identifiers unchanged.
Typography QA: Use correct English spelling, punctuation, capitalization, and natural phrase-based line breaks. Treat non-English Text (verbatim) content as source copy to localize unless the recipe is explicitly multilingual. Do not add filler text.
Use case: social-media-content
Asset type: vertical 3:4 phone photograph for a neighborhood restaurant story
Goal: make a freshly served [DISH] feel immediate, honest and locally grounded rather than studio-perfect
Scene/backdrop: small wooden table near a window at lunchtime, used napkin, water glass, one blurred neighboring table
Subject: one serving of [DISH], steam or condensation only when physically appropriate, small imperfections retained
Camera language: modern phone main camera, 24 mm equivalent, close focus, mild computational sharpening, realistic phone depth of field
Composition/framing: slightly off-center overhead three-quarter angle, cropped utensils at edge, natural handheld perspective
Lighting/mood: soft window light with weak ceiling-light contamination, appetizing but unretouched
Color: accurate food color, neutral white balance, restrained saturation
Constraints: dish structure must match supplied description; no extra garnish; believable portion; no brand marks; no impossible blur
Avoid: glossy stock-photo styling, floating ingredients, excessive steam, spotless unused table, fake film border, watermark
Use it for: photograph authorized [PRODUCT] with a visual system that can be repeated across future SKUs
Target language: English (en)
Localization directive: Convert every user-editable variable and non-English quoted example into natural English before rendering. Preserve deliberately multilingual copy only when the recipe explicitly requires it. Keep authorized brand and product identifiers unchanged.
Typography QA: Use correct English spelling, punctuation, capitalization, and natural phrase-based line breaks. Treat non-English Text (verbatim) content as source copy to localize unless the recipe is explicitly multilingual. Do not add filler text.
Use case: ecommerce-product
Asset type: 4:5 premium product campaign image plus reproducible lighting specification
Goal: photograph authorized [PRODUCT] with a visual system that can be repeated across future SKUs
Product lock: exact geometry, cap, label hierarchy, material, color and supplied brand assets
Set: matte limestone sweep, one translucent acrylic plane behind product, no unrelated props
Camera language: 85 mm product lens, camera height at product midpoint, minimal perspective distortion, full label sharp
Lighting: key softbox 45° camera left and 30° above; narrow strip rim camera right; white bounce below; black flag behind left edge
Composition: product at 45% horizontal position, 30% negative space for approved copy, shadow falls back-right
Color palette: [BRAND COLORS] with neutral stone and one controlled accent
Text (verbatim): "[HEADLINE]", "[APPROVED BENEFIT]"
Constraints: exact product; lighting directions visibly consistent; two copy lines once each; no invented certification or claim
Avoid: altered packaging, extra products, liquid splash unless supplied, clipped highlights, random props, watermark
Use it for: establish a quiet late-night turning point outside a rural bus stop without showing recognizable intellectual property
Target language: English (en)
Localization directive: Convert every user-editable variable and non-English quoted example into natural English before rendering. Preserve deliberately multilingual copy only when the recipe explicitly requires it. Keep authorized brand and product identifiers unchanged.
Typography QA: Use correct English spelling, punctuation, capitalization, and natural phrase-based line breaks. Treat non-English Text (verbatim) content as source copy to localize unless the recipe is explicitly multilingual. Do not add filler text.
Use case: cinematic-concept
Asset type: 2.39:1 cinematic establishing frame
Goal: establish a quiet late-night turning point outside a rural bus stop without showing recognizable intellectual property
Scene/backdrop: wet two-lane road, modest glass shelter, one timetable panel with no readable text, distant farmhouse lights, low cloud
Subjects: one small adult silhouette waiting with a canvas bag; an approaching bus visible only as soft distant headlights
Camera language: 35 mm anamorphic-inspired framing without fake brand artifacts, tripod height 1.4 m, deep environmental composition
Exposure logic: practical shelter light exposes the figure edge; headlights bloom mildly through mist; sky retains low-detail cloud texture
Color grade: sodium amber, deep blue-gray, desaturated roadside green
Continuity anchors: shelter shape, bag, road markings and weather must remain fixed for later shots
Constraints: safe roadside positioning; plausible light falloff; no readable route number; no franchise reference
Avoid: neon cyberpunk, giant lens flare, crushed blacks, daylight-bright subject, extra vehicles, watermark
Use it for: create a believable analog portrait session with one authorized [SUBJECT] while preserving identity across every frame
Target language: English (en)
Localization directive: Convert every user-editable variable and non-English quoted example into natural English before rendering. Preserve deliberately multilingual copy only when the recipe explicitly requires it. Keep authorized brand and product identifiers unchanged.
Typography QA: Use correct English spelling, punctuation, capitalization, and natural phrase-based line breaks. Treat non-English Text (verbatim) content as source copy to localize unless the recipe is explicitly multilingual. Do not add filler text.
Use case: portrait-series
Asset type: 4×3 contact sheet of twelve distinct frames
Goal: create a believable analog portrait session with one authorized [SUBJECT] while preserving identity across every frame
Identity lock: exact face, age, hair, skin tone, body proportions and [OUTFIT]
Location: plain daylight studio with stool, white wall and one window; no set changes
Shot list: three headshots, three waist-up gestures, three seated variations, two motion-blur experiments, one quiet profile
Camera language: 35 mm color-negative film, 50 mm lens, exposure variation within one stop, natural frame-to-frame focus differences
Sheet design: twelve equal frames with thin black gutters; small frame numbers only "01" through "12"
Texture: real film grain, minor dust and color variation; no fake severe scratches
Constraints: exactly twelve frames; same person and outfit; unique pose in each; numbers correctly mapped
Avoid: face drift, duplicated frame, wardrobe changes, extreme aging, fake sprocket-brand text, watermark
Use it for: document a compact [ROOM TYPE] with accurate geometry, exterior view and material color
Target language: English (en)
Localization directive: Convert every user-editable variable and non-English quoted example into natural English before rendering. Preserve deliberately multilingual copy only when the recipe explicitly requires it. Keep authorized brand and product identifiers unchanged.
Typography QA: Use correct English spelling, punctuation, capitalization, and natural phrase-based line breaks. Treat non-English Text (verbatim) content as source copy to localize unless the recipe is explicitly multilingual. Do not add filler text.
Use case: architecture-photography
Asset type: 4:3 professional interior photograph
Goal: document a compact [ROOM TYPE] with accurate geometry, exterior view and material color
Space lock: preserve supplied floor plan, window count, ceiling height, built-ins and furniture positions
Camera language: 24 mm tilt-shift look, camera level at 1.35 m, verticals parallel, moderate field of view without stretched corners
Exposure: naturally blended interior and window exposure; window view is present but secondary; no HDR glow
Lighting: existing daylight plus practical fixtures at believable intensity; contact shadows remain visible
Materials: distinguish oak, lime plaster, linen, brushed metal and stone through surface response
Composition: two-point view with foreground depth, clear circulation path, no staged person
Constraints: no architectural redesign; plausible scale; all lights attached; accurate reflections and perspective
Avoid: ultra-wide distortion, floating furniture, repeated decor, oversaturated sky, sterile CGI finish, watermark
Use it for: capture an authorized amateur [SPORT] athlete at the decisive moment while keeping movement anatomically plausible
Target language: English (en)
Localization directive: Convert every user-editable variable and non-English quoted example into natural English before rendering. Preserve deliberately multilingual copy only when the recipe explicitly requires it. Keep authorized brand and product identifiers unchanged.
Typography QA: Use correct English spelling, punctuation, capitalization, and natural phrase-based line breaks. Treat non-English Text (verbatim) content as source copy to localize unless the recipe is explicitly multilingual. Do not add filler text.
Use case: sports-editorial
Asset type: 16:9 editorial action photograph
Goal: capture an authorized amateur [SPORT] athlete at the decisive moment while keeping movement anatomically plausible
Subject lock: supplied athlete identity, uniform construction, equipment and bib number
Action: [PRECISE SPORT MOVEMENT] at the biomechanically correct phase; weight, joint angles and equipment contact must agree
Camera language: 300 mm sideline lens, low viewpoint, 1/2000 s freeze, subject sharp, crowd softly compressed
Environment: realistic community venue, lane or field markings correct for the sport, no real sponsor signage
Lighting: late-afternoon side light, moderate contrast, natural sweat and fabric movement
Composition: space in direction of motion, horizon level, no cropped hands or equipment
Constraints: correct limb count, grip, ball/equipment geometry and rules context; authorized athlete only
Avoid: impossible pose, fused equipment, superhero physique, random text, duplicated players, watermark
Use it for: create a conservative restored version on the right while documenting the visual difference
Target language: English (en)
Localization directive: Convert every user-editable variable and non-English quoted example into natural English before rendering. Preserve deliberately multilingual copy only when the recipe explicitly requires it. Keep authorized brand and product identifiers unchanged.
Typography QA: Use correct English spelling, punctuation, capitalization, and natural phrase-based line breaks. Treat non-English Text (verbatim) content as source copy to localize unless the recipe is explicitly multilingual. Do not add filler text.
Use case: image-restoration
Asset type: side-by-side restoration proof at the source image ratio
Input images: Image 1 is an authorized faded family photograph; left panel must show it unchanged
Goal: create a conservative restored version on the right while documenting the visual difference
Restoration: neutralize color cast, recover moderate contrast, reduce dust and scratches, repair only clearly damaged areas
Identity lock: preserve every person’s face, age, expression, hairstyle, body shape, clothing and position
Historical lock: preserve original objects, signage, architecture, film softness and period color plausibility
Layout: equal before/after panels labeled only "BEFORE" and "RESTORED"
Constraints: do not invent missing facial detail; do not add people or objects; no modernized clothing; retain natural grain
Avoid: face enhancement, recolored fantasy palette, excessive sharpening, AI beauty treatment, fabricated background, watermark