morpheus-fashion-design（墨菲斯时尚设计）

Morpheus Fashion Design

Generate professional fashion/product advertising images using ComfyDeploy's Morpheus Fashion Design workflow.

⚠️ CRITICAL RULE: NEVER USE AUTO VALUES

Configuration packs MUST NEVER be left on auto or AUTO.

auto = empty values = neutral, boring images with no creative direction.

The pack options listed below are suggestions/ideas, but you can send custom values that better fit the brief. The goal is to select the best possible configuration to represent the image needed for the brief.

Pack Selection Guidelines

For EVERY generation, thoughtfully select values based on the creative brief:

Example: Oakley Snowboarding Campaign

style_pack = "cinematic_realism"  # NOT auto - sports action needs energy
camera_pack = "sony_a1"            # Fast sports camera
lens_pack = "wide_distortion_controlled"  # Capture the action
lighting_pack = "golden_hour_backlit"     # Alpine dramatic lighting
pose_discipline_pack = "sport_in_motion"  # Rider in action
time_weather_pack = "golden_hour_clear"   # Mountain conditions

Custom Values

If none of the preset options fit, you can write your own value as a descriptive string:

lighting_pack = "harsh alpine midday sun reflecting off fresh powder"
environment_pack = "snowpark with metal rails and pristine packed snow"

Overview

Morpheus Fashion Design is a comprehensive AI workflow for creating high-quality commercial photography with:

• Product integration
• Model face consistency
• Professional lighting and camera settings
• Brand-aligned creative direction

API Details

Endpoint: https://api.comfydeploy.com/api/run/deployment/queue Deployment ID: 1e16994d-da67-4f30-9ade-250f964b2abc (production)

Required Inputs

Images (must be URLs)

1. product - Product image URL (the item being advertised)
2. model - Model face reference (frontal face photo)
3. logo - Brand logo (optional, use placeholder if not needed)

🎭 Model Catalog

A curated catalog of 114 AI-generated model references is available for use when no specific model is provided.

Repository

GitHub: https://github.com/PauldeLavallaz/model_management

⚠️ PRIORITY: User-provided model ALWAYS wins

If the user attaches/provides a model image → use that image directly. The catalog is ONLY for when no model is specified.

Setup (First Time Installation)

# Clone the catalog to your workspace
git clone https://github.com/PauldeLavallaz/model_management.git models-catalog

Update Catalog

cd models-catalog && git pull

Local Path (if already cloned)

~/clawd/models-catalog/catalog/images/

Catalog Structure

models-catalog/
└── catalog/
    ├── catalog.json      # Full metadata for all models
    └── images/           # Model reference photos (model_01.jpg - model_114.jpg)

Using the Catalog

Priority order for model selection:

1. User provides model image → Use that image directly
2. User describes desired model → Search catalog and select best match
3. No specification → Select appropriate model based on campaign brief

Searching the Catalog

# List all models with basic info
cat models-catalog/catalog/catalog.json | jq '[.talents[] | {id, name, gender, ethnicity, tags: .tags[0:2]}]'

# Find models by ethnicity
cat models-catalog/catalog/catalog.json | jq '[.talents[] | select(.ethnicity == "hispanic") | {id, name, description}]'

# Find models by tag
cat models-catalog/catalog/catalog.json | jq '[.talents[] | select(.tags[] == "commercial") | {id, name, ethnicity}]'

# Find models by gender
cat models-catalog/catalog/catalog.json | jq '[.talents[] | select(.gender == "male") | {id, name, ethnicity}]'

Model Attributes

Each model entry includes:

• id: Unique identifier (model_01, model_02, etc.)
• name: Model name
• gender: female, male, non-binary
• ethnicity: african, asian, caucasian, hispanic, mixed, etc.
• age_group: young_adult, adult, mature
• tags: editorial, commercial, beauty, lifestyle, avant-garde, etc.
• description: Detailed description of look and best uses
• image_path: Path to reference image

Example: Selecting a Model

# For an Argentine campo/gaucho campaign, find hispanic females with commercial tags:
cat models-catalog/catalog/catalog.json | jq '[.talents[] | select(.ethnicity == "hispanic" and .gender == "female" and (.tags[] == "commercial" or .tags[] == "lifestyle")) | {id, name, description}]'

# Then use the selected model:
--model "models-catalog/catalog/images/model_08.jpg"

Creative Brief

1. brief - Detailed campaign description including:

- Scene/location description - Model pose and action - Product placement and interaction - Lighting and mood - Camera technique - Visual style

1. target - Target audience description including:

- Demographics - Psychographics - Interests and lifestyle

Configuration Packs

Workflow Process

1. Receive request with brand/product info
2. Design brief aligned with brand identity and campaign goals
3. Define target audience with demographics and psychographics
4. Prepare images:

- Download product image(s) - Download/find model reference (frontal face) - Upload to ComfyDeploy storage to get URLs

1. Select packs based on creative direction
2. Submit job and poll for completion
3. Deliver results

Usage

uv run ~/.clawdbot/skills/morpheus-fashion-design/scripts/generate.py \
  --product "path/to/product.jpg" \
  --model "path/to/model-face.jpg" \
  --brief "Campaign brief text..." \
  --target "Target audience description..." \
  --aspect-ratio "4:5" \
  --style-pack "street_authentic" \
  --output "output-filename.png"

Example Brief (Franuí Carnaval)

La campaña Franuí Carnaval captura el espíritu festivo y la alegría del carnaval brasileño
en Copacabana. Una mujer afrobrasileña baila en medio de la multitud, sosteniendo el
producto Franuí Milk hacia la cámara en un gesto espontáneo y celebratorio. La escena
está llena de confeti, movimiento y energía. La fotografía adopta un estilo documental
con motion blur intencional, ángulo bajo que empodera al sujeto, y el producto como
elemento hero en primer plano. La luz es natural de día tropical, cálida y vibrante.

Example Target (Franuí Carnaval)

Jóvenes adultos 18-35, principalmente mujeres pero inclusivo, que celebran la vida,
la música y los momentos compartidos. Consumidores de experiencias premium que buscan
productos que se integren naturalmente a sus momentos de disfrute. Activos en redes
sociales, valoran la autenticidad y la conexión cultural. Mercado: Brasil y LATAM.

Important Notes

Studio Override

The workflow has an automatic studio_override that activates when no location reference is provided. This will use a white cyclorama background regardless of the brief description.

To get environmental backgrounds:

1. Provide a location_ref image, OR
2. Set environment_pack to a specific environment (e.g., beach_minimal, street_crosswalk)

Priority Hierarchy

The system follows this priority:

1. Talent (identity preservation) >
2. Garments (product fidelity) >
3. Fit > Pose > Style > Location > Branding

API Key

DO NOT pass the API key via parameter. Leave it empty.

The API key is already configured in ComfyDeploy. Passing --api-key will cause authentication errors.

Troubleshooting

Imagen negra o vacía

Si la imagen generada sale completamente negra o vacía, es el filtro de moderación de Google/Gemini. Causas comunes:

• Se pidió una persona famosa o celebridad
• Contenido considerado sensible por el modelo
• La combinación de prompt + imágenes fue rechazada por políticas de contenido

Solución: Modificar el prompt para evitar referencias a personas reales/famosas, o cambiar elementos que puedan activar el filtro.