morpheus-fashion-design

---
name: morpheus-fashion-design
description: |
  Generate professional advertising images with AI models holding/wearing products.
  
  ✅ USE WHEN:
  - Need a person/model in the image WITH a product
  - Creating fashion ads, product campaigns, commercial photography
  - Want consistent model face across multiple shots
  - Need professional lighting/camera simulation
  - Input: product image + model reference (or catalog)
  
  ❌ DON'T USE WHEN:
  - Just editing/modifying an existing image → use nano-banana-pro
  - Product-only shot without a person → use nano-banana-pro
  - Already have the hero image, need variations → use multishot-ugc
  - Need video, not image → use veed-ugc after generating image
  - URL-based product fetch with brand profile → use ad-ready instead
  
  OUTPUT: Single high-quality PNG image (2K-4K resolution)
---

# 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:

| Pack | How to Choose |
|------|---------------|
| `style_pack` | Match brand personality: luxury→`premium_restraint`, sports→`cinematic_realism`, street→`street_authentic` |
| `camera_pack` | What camera would a real photographer use? Sports→`sony_a1`, editorial→`hasselblad_x2d`, street→`leica_m6` |
| `lens_pack` | Portrait compression? Anamorphic? Wide? Match the shot type and mood |
| `lighting_pack` | What's described in the brief? Golden hour? Studio? Natural window? Choose accordingly |
| `pose_discipline_pack` | What's the model doing? Sport action→`sport_in_motion`, commercial→`commercial_front_facing` |
| `film_texture_pack` | Warm editorial→`kodak_portra_400`, cinematic→`kodak_vision3_500t`, clean digital→`digital_clean_no_emulation` |
| `environment_pack` | Match brief location: beach→`beach_minimal`, urban→`urban_glass_steel`, nature→provide location_ref image |
| `color_science_pack` | Warm tones? Cool? Cinematic contrast? Select based on mood |
| `time_weather_pack` | When does the scene happen? Golden hour? Midday? Overcast? |

### Example: Oakley Snowboarding Campaign
```python
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:
```python
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)
```bash
# Clone the catalog to your workspace
git clone https://github.com/PauldeLavallaz/model_management.git models-catalog
```

### Update Catalog
```bash
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
```bash
# 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
```bash
# 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

2. **target** - Target audience description including:
   - Demographics
   - Psychographics
   - Interests and lifestyle

### Configuration Packs

| Pack | Options |
|------|---------|
| `style_pack` | auto, premium_restraint, editorial_precision, cinematic_realism, cinematic_memory, campaign_hero, product_truth, clean_commercial, street_authentic, archive_fashion, experimental_authorial |
| `shot_pack` | auto, full_body_wide, medium_shot, close_up, low_angle_hero, three_quarter, waist_up, etc. |
| `camera_pack` | auto, arri_alexa35, canon_r5, hasselblad_x2d, leica_m6, sony_a1, etc. |
| `lens_pack` | auto, cooke_anamorphic_i_50, leica_noctilux_50, zeiss_otus_55, etc. |
| `lighting_pack` | auto, golden_hour_backlit, natural_window, studio_three_point, etc. |
| `pose_discipline_pack` | auto, commercial_front_facing, street_style_candid_walk, sport_in_motion, etc. |
| `film_texture_pack` | auto, kodak_portra_400, fujifilm_velvia_50, digital_clean_no_emulation, etc. |
| `color_science_pack` | auto, neutral_premium_clean, warm_golden_editorial, cinematic_low_contrast, etc. |
| `environment_pack` | AUTO, beach_minimal, urban_glass_steel, street_crosswalk, etc. |
| `time_weather_pack` | auto, golden_hour_clear, bright_midday_sun, overcast_winter_daylight, etc. |
| `branding_pack` | logo_none, logo_discreet_lower, logo_top_corner, logo_center_watermark, logo_integrated |
| `intent` | auto, awareness, consideration, conversion, retention |
| `aspect_ratio` | 9:16, 16:9, 1:1, 4:5, 5:4, 3:4, 4:3 |

## 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
5. **Select packs** based on creative direction
6. **Submit job** and poll for completion
7. **Deliver results**

## Usage

```bash
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.