🔥
10d • 💸 AI Ideas
⭕️//::: 𝗦𝗲𝗲𝗱𝗮𝗻𝗰𝗲 𝟮.𝟱 ~ 𝗧𝘂𝗿𝗻 𝗢𝗻𝗲 𝗜𝗺𝗮𝗴𝗲 𝗜𝗻𝘁𝗼 𝗮 𝗩𝗶𝗱𝗲𝗼 𝗖𝗵𝗮𝗿𝗮𝗰𝘁𝗲𝗿 𝗦𝗵𝗲𝗲𝘁
𝗛𝗲𝘆 𝗙𝗮𝗺, 👊🏾:::
I was up early this morning working on a project, and while I was at it, I started playing around with the NEW Seedance 2.5.
... 𝘁𝗵𝗶𝘀 𝘁𝗵𝗶𝗻𝗴 𝗶𝘀 𝗮𝗻 𝗮𝗯𝘀𝗼𝗹𝘂𝘁𝗲 𝗯𝗲𝗮𝘀𝘁. 🔥
While testing, I came up with a really solid prompt that can help you turn a single image into a clean video character sheet.
So if you’re building AI videos and you want a stronger visual reference for your character, this should help a lot.
Just upload the image you want to use, paste in the prompt below, and let it do its thing.
Use it.
Remix it.
Share it.
Run with it.
I’m dropping it here for the community.
𝗖𝗼𝗽𝘆 𝗮𝗻𝗱 𝗽𝗮𝘀𝘁𝗲 𝗳𝗿𝗼𝗺 𝗵𝗲𝗿𝗲 𝗱𝗼𝘄𝗻:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CHARACTER REFERENCE SHEET — LOCKED LAYOUT TEMPLATE
SOURCE: [attached image]. This is the character. Reproduce her exactly —
same face, same proportions, same outfit, same colors, same art style,
same rendering technique and line quality in every single element below.
CANVAS: 4:3 landscape. Flat warm off-white background (#F4F1EC). No
texture, no gradient, no outer border or frame. Generous margins.
═══ ZONE A — TURNAROUND ═══
Position: upper LEFT. Occupies ~70% of canvas width, top ~62% of height.
Header centered at top of zone: the word "TURNAROUND" in thin all-caps
letterspaced sans-serif, warm dark grey. Flanked left and right by a
single thin horizontal hairline rule, each rule terminating in a small
filled dot at the outer end.
Below the header: FOUR full-body figures in ONE horizontal row, evenly
spaced, equal height, all feet resting on the same invisible baseline,
each with a soft contact shadow. Left to right:
1. FRONT — facing viewer straight on
2. THREE-QUARTER — body rotated ~45°
3. SIDE — full 90° profile
4. BACK — rear view, no face visible
Identical pose, identical outfit, identical accessories held in the same
hand across all four. Small all-caps label centered beneath each figure:
FRONT / THREE-QUARTER / SIDE / BACK
═══ ZONE B — FACIAL EXPRESSIONS ═══
Position: RIGHT column, ~30% of canvas width, running the full height of
the upper section, aligned to the same top margin as Zone A.
Header at top of zone: "FACIAL EXPRESSIONS" in the same type treatment,
with the same hairline rule and dot terminator on its left side.
Below it: SIX head-and-shoulders portraits, cropped just below the
collar, all at identical scale. Arranged in a perfectly even 2-column ×
3-row grid — every row holds exactly two portraits, no row is centered
or offset, no portrait stands alone:
Row 1: NEUTRAL | HAPPY
Row 2: CONFIDENT | SURPRISED
Row 3: THOUGHTFUL | ANGRY
Small all-caps label centered beneath each portrait.
═══ ZONE C — DETAIL CALL-OUTS ═══
Position: LOWER LEFT, directly beneath Zone A. Same ~70% width, bottom
~28% of canvas height.
Header centered at top of zone: "DETAIL CALL-OUTS" in the same type
treatment, flanked by hairline rules with dot terminators.
Below it: THREE circular vignettes in one horizontal row, evenly spaced,
identical diameter, each ringed by a thin metallic-gold hairline circle.
Each contains a tight close-up crop of a distinctive costume or accessory
element belonging to THIS character — choose the three most defining
elements (e.g. hairstyle, a garment construction detail, a grouping of
carried accessories). Small all-caps label centered beneath each circle.
═══ TYPOGRAPHY ═══
One thin sans-serif throughout. All caps. Wide letterspacing. Warm dark
grey-brown. Section headers ~1.5× the size of item labels. No other text
anywhere on the sheet.
═══ FIDELITY LOCK ═══
Facial features, skin tone, hair shape and volume, body proportions,
garment cut and construction, seam and button placement, accessory
color and hardware must be pixel-consistent with the source image and
with each other across all thirteen renders on this sheet.
═══ DO NOT ═══
No identity drift between views. No duplicated or mirrored poses. No
missing views. No outfit variation. No cropped feet, hands, or heads.
No stacked or overlapping figures. No extra characters. No logos,
watermarks, signatures, page numbers, color swatches, measurement lines,
annotation arrows, or grid guides. No drop shadows on text. No panel
borders or boxes around zones.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━
If you use it, post up what you make so we can all see what you cooked up.
✌🏾🖤
~𝗔𝗱𝗿𝗶𝗮𝗻
𝗗𝗼𝗻'𝘁 𝗷𝘂𝘀𝘁 𝗹𝗲𝗮𝗿𝗻, 𝗹𝗲𝗮𝗱.
//
10
13 comments
⭕️//::: 𝗦𝗲𝗲𝗱𝗮𝗻𝗰𝗲 𝟮.𝟱 ~ 𝗧𝘂𝗿𝗻 𝗢𝗻𝗲 𝗜𝗺𝗮𝗴𝗲 𝗜𝗻𝘁𝗼 𝗮 𝗩𝗶𝗱𝗲𝗼 𝗖𝗵𝗮𝗿𝗮𝗰𝘁𝗲𝗿 𝗦𝗵𝗲𝗲𝘁
skool.com/ai-driven-business-summit
Build a smarter business with AI. Monthly mentors, AI tools, and summits on Skool for midlife women entrepreneurs. Next Summit: October 5-7, 2026.
Leaderboard (30-day)
1
+1281
2
🔥
+374
3
+328
4
🔥
+319
5
+256
Powered by