Cinematic AI Presets: 8 Visual Worlds from a Single Selfie ⚡
The one-command portrait architecture created by Hancybox (@hancybox). How to transform any standard phone selfie into 8 high-production editorial scenes while strictly locking your facial proportions, authentic skin texture, and outfit styling.
⚡ TL;DR
- The Core Rule: "Preserve the person, not the face in the preset." Your selfie supplies 100% of identity; the preset supplies 100% of composition, lighting, and cinematic mood.
- Zero Generic AI Faces: Enforces an explicit anti-beautification directive that forbids plastic skin smoothing, face-blending, ethnicity alteration, and symmetry flattening.
- 8 Production Aesthetics: Access 8 distinct visual setups (
/Reign,/Afterlife,/Overlord,/Void,/Untouchable,/Havoc,/Hunter,/Nocturne) with single slash commands. - Universal Portability: Ready-to-use in ChatGPT (Vision + Image Generation), Midjourney (via
--crefcharacter weight), and Flux.1.
The AI Photography Paradox: Great Scenes, Fake Faces
Every creator who has tried generating editorial portraits of themselves with AI hits the exact same wall: either the lighting and film grain look incredible but the face looks like a generic 22-year-old Instagram model, or the face resembles you but is pasted awkwardly onto flat, lifeless stock lighting.
The Cinematic Preset Kit (developed by digital artist Hancybox — @hancybox across Instagram, YouTube, and TikTok) completely cracks this problem with an elegant architectural separation:
1. Identity Lock (User Selfie): Supplies bone structure, eye shape, nose bridge, jawline, skin tone, natural texture, hairstyle, visible facial hair, accessories, and outfit.
2. Visual Lock (Preset Reference): Supplies pose, focal length, camera perspective, lighting ratios, color grading (deep cyan/blue dominance), shadow occlusion, motion blur, and film imperfections.
Never transfer the preset subject's face to the user. Instead, recreate the preset's cinematic atmosphere with the user's authentic, recognizable identity.
The Master Identity Rules (Read Before Generating)
Whenever you run any of the 8 presets below in ChatGPT, Midjourney, or Flux, pair your selfie with this master instruction block to guarantee true facial fidelity:
The 8 Cinematic Preset Worlds & Commands
Upload your selfie, attach the master identity instructions, and run any of these 8 commands to generate your new scene:
/Reign — Dominant Close Portrait
Aesthetic: Dominant, low-key close crop with dramatic edge rim lighting and an intense, unwavering eye line. Soft anamorphic lens flare sweeps across the lower foreground while the subject emerges from pitch shadow.
/ReignVisual Code: Low-key lighting, sharp rim highlights, 85mm portrait crop, intense gaze, zero facial smoothing.
/Afterlife — Surreal Flowers & Butterflies
Aesthetic: Dreamlike flower-field composition with soft cinematic imperfection. The subject reclines naturally as luminous butterflies and blurred foreground flora create deep spatial layering against a moody night atmosphere.
/AfterlifeVisual Code: Surreal botanical depth, floating bioluminescent butterflies, soft aperture blur, dreamcore color grading.
/Overlord — Giant Defocused Eyes & Blue Shadow
Aesthetic: Centered psychological cinematic portrait. Looming in the background behind the subject are massive, blurred, glowing blue eyes that establish an ominous, powerful presence, while deep half-shadow split lighting sculpts the user's face.
/OverlordVisual Code: Psychological double-exposure depth, split chiaroscuro lighting, electric blue ocular background, cyberpunk thriller tone.
/Void — Blue Veil & Silhouette
Aesthetic: Editorial abstract portrait with foreground light obstruction and sheer cyan veiling. Deep silhouettes and negative space frame a mysterious, magnetic direct gaze.
/VoidVisual Code: Foreground light refraction, deep silhouette occlusion, negative space balance, moody indigo color grade.
/Untouchable — Focal Subject in a Blue Crowd
Aesthetic: The protagonist amidst the noise. The user is embedded in a dense, dark crowd where surrounding faces blur into the shadows, leaving the user illuminated as the sharp, calm epicenter of the frame.
/UntouchableVisual Code: Shallow depth of field crowd separation, isolated blue key light, protagonist framing, moody urban atmosphere.
/Havoc — Seated Chaos & Flying Cards
Aesthetic: Controlled anarchy. The user is seated in a relaxed, confident reclined position while playing cards and fiery spark trails erupt through the foreground in slow-shutter motion blur. High contrast between deep navy shadows and bright orange ember streaks.
/HavocVisual Code: Dynamic particle blur, flying cards in motion, teal-and-orange cinematic split, nonchalant seated body language.
/Hunter — Archery Portrait & Blue Backlight
Aesthetic: Focused athleticism. The bow forms a striking diagonal leading line, with crisp blue rim lighting highlighting the archer's shoulder and eye line as they draw. Moody, cinematic, and intense without looking like a costume.
/HunterVisual Code: Sporting bow prop composition, electric blue rim backlight, diagonal tension leading line, laser focus eye contact.
/Nocturne — Portrait Through Light Bands
Aesthetic: Neo-noir venetian blinds. Razor-sharp horizontal light and shadow bands dissect the face, creating geometric mystery while leaving the eyes burning brightly through the darkness.
/NocturneVisual Code: Venetian blind gobos, horizontal hard light slicing, pitch shadow negative space, intense neo-noir thriller tone.
How to Apply Custom Modifiers & Overrides
One of the strongest features of the Hancybox kit is single-dimension overrides. If you want to tweak one detail (like grain, lighting intensity, or color balance) without breaking the rest of the preset, append your modifier directly to the command:
Practical Command Override Examples:
/Reign no grain— Removes analog 35mm grain for an ultra-clean, modern digital cinema finish./Void face slightly brighter— Lifts shadow values across the nose and cheekbone for higher identity recognition in dark rooms./Havoc 90% blue 10% orange— Restricts the warm fire embers to subtle, pinpoint accents for a cooler overall image./Untouchable no crossed arms— Replaces defensive body language with a natural walking-forward or hands-in-pockets posture.
How to Run Across Different AI Image Engines
Upload your selfie directly in chat. Paste the System Prompt • Identity Lock block first, followed by your chosen preset command (e.g., /Havoc). GPT-4o analyzes your facial structure and guides DALL-E/Image Gen to render you directly inside the scene.
Upload your selfie to Discord to obtain its URL. Paste the preset prompt text and append --cref <YOUR_SELFIE_URL> --cw 80. The --cw 80 parameter locks your face and hair structure while allowing Midjourney to adapt lighting and wardrobe to the preset's cinematic atmosphere.
Load Flux.1 Dev with PuLID (Pure Lightning Identity) in ComfyUI or Fal.ai. Feed your selfie into the PuLID node (weight: 0.82) and paste the exact preset prompt into the positive prompt conditioning. Flux produces photorealistic film grain and micro-textures without plastic smoothing.