Consistent Characters
Your videos and images can now use the same character for brand consistency
If you happened to find your way here from my YouTube on creating consistent characters in Gemini Veo3... Welcome. Glad you stopped by.
Use this prompt to have ChatGPT fully describe the avatar image you have created and uploaded to the chat.
Once you run this prompt, you’ll get a detailed analysis of your avatar in bullet list format. Copy the entire ChatGPT result and paste into a new prompt. Then, use the “Convert ChatGPTs result to JSON” prompt (below the character description prompt at the bottom of the page).
---Here’s the Character Description prompt---
CONTEXT: I am looking to create videos with text-to-video solutions like Google’s Gemini VEO3 and Runway ML. Part of the challenge of this process is the ability to create consistent characters. I will load a photo into ChatGPT and have ChatGPT fully describe the person in the image based on the list of characteristics below. This way, VEO3 can recreate the same person for each new scene. This will require extensive detail on all aspects of the subject.
ROLE You are a highly renowned forensic facial examiner and medical aesthetician with more than 20 years of experience analyzing and describing facial features.
Here is the list of characteristics for you to reference when completely describing the person in the photo
1. Identity & Demographic Cues
Apparent age range
Gender presentation
Ancestry / ethnic cues (skin undertone, facial bone structure, hair pattern)
Approximate height and build (short, tall, slim, broad, etc.)
Typical posture (upright, relaxed, slouched)
2. Facial Structure
Head shape (oval, round, square, heart, diamond)
Forehead height and slope
Cheekbone prominence and angle
Jawline contour (sharp, soft, rounded)
Chin shape (cleft, pointed, broad)
3. Hair
Color (base tone, highlights, lowlights)
Length and cut (buzzed, bob, shoulder-length, waist-length)
Texture (straight, wavy, curly, coily)
Hairline pattern and parting
Styling details (braids, ponytail, loose, gelled)
4. Eyes & Brows
Eye shape (almond, hooded, deep-set, monolid)
Iris color and pattern (hazel with golden flecks, dark brown, ice blue, etc.)
Sclera tint (bright white, slightly veined)
Eyebrow shape, thickness, grooming (arched, bushy, thin, natural)
Eyelash length and fullness
5. Nose, Mouth & Ears
Nose bridge height, width, tip shape, nostril flare
Lip fullness, cupid’s bow definition, natural color tint
Teeth visibility (gap, straight, slight overlap)
Ear size, lobe attachment, piercings
6. Skin
Base tone (light olive, deep mahogany, medium tan)
Surface texture (smooth, fine lines, visible pores)
Undertone (cool, neutral, warm)
Freckles, moles, scars, birthmarks
Makeup or bare skin (foundation finish, blush placement, highlights)
7. Facial & Body Hair
Beard or mustache style, length, density, edge lines
Sideburn length and fade
Arm, leg, or chest hair visibility (for uncovered areas)
8. Body Proportions
Shoulder width versus hip width
Torso length
Limb length (long-legged, short-armed, proportional)
Muscle definition (athletic, average, soft)
9. Clothing
Garment type and fit (tailored suit, loose hoodie, cropped denim jacket)
Color palette and patterns
Fabric texture (wool tweed, cotton jersey, silk satin)
Layering and closures (single-breasted, zip-front, button-down)
Condition and style cues (pressed, distressed, vintage, modern)
10. Accessories
Eyewear (frame shape, material, lens tint)
Jewelry (ear studs, hoop earrings, layered necklaces, wristwatch)
Headwear (cap, beanie, fedora)
Bags, belts, gloves, scarves
Tech items (earbuds, smartwatch)
11. Distinguishing Marks & Modifiers
Tattoos (placement, style, color, size)
Piercings beyond ears (nose ring, eyebrow bar)
Medical devices (hearing aid, walking cane)
Temporary elements (bandages, stage makeup)
12. Posture, Gesture & Expression
Weight distribution (leaning on left hip, balanced)
Hand position (arms crossed, hands in pockets)
Facial expression baseline (neutral, friendly smile, intense glare)
Micro-expressions (raised brow, slight smirk)
13. Lighting & Color Grading
Key light direction and quality (soft window light from camera left, hard top light)
Fill and rim presence
Color temperature (warm 3200 K, cool 5500 K)
Shadow depth and softness
Overall grade (high-contrast noir, pastel low-contrast, cinematic teal-orange)
14. Camera & Framing
Lens focal length (35 mm environmental, 85 mm portrait, 16 mm wide)
Aperture / depth of field (shallow bokeh, deep focus)
Angle (eye level, low-angle hero, high-angle)
Crop (full body, three-quarter, tight headshot)
Motion blur expectations for moving shots
15. Environment & Context (if visible)
Setting type (urban street, studio seamless, forest clearing)
Background detail level (clean backdrop, textured brick wall)
Props in hand or nearby (coffee cup, guitar, clipboard)
Color interplay between subject and setting
16. Movement & Action Cues (for animation continuity)
Typical gait or walk cycle (brisk, casual, stooped)
Signature gestures (adjusting glasses, brushing hair back)
Interaction style with objects or other characters
17. Stylistic Tags & References
Era or subculture influence (’90s grunge, cyberpunk, 1950s classic)
Cinematic or photographic references (film grain Kodak Portra 400, glossy Vogue editorial)
Descriptive keywords used consistently across prompts (e.g., “Jordan K., shaved-side undercut, emerald eyes, freckled nose”)
---Convert ChatGPT’s result to JSON with this prompt---
Please write this description in JSON format. Don’t elaborate or create any type if story. Simply use the categories provided to define the person’s face in JSON inclusive of the descriptive details.