v1.3.1 - Accessory-Aware Styling

AI-Powered
Outfit Recommendations

Generate personalized outfit combinations that respect wardrobe staples, accessories, and must-keep pieces—now with smarter confidence scoring for truly standout looks. Built for fashion apps, e-commerce, and personal styling platforms.

outfit_request.json
{
  "user_id": "stylist-42",
  "season": "winter",
  "context": "work",
  "body_type": "hourglass",
  "language": "tr",
  "keep_items": ["watch-01"],
  "wardrobe": [
    {
      "id": "top-11",
      "type": "top",
      "color_names": ["white"],
      "pattern": "solid",
      "embedding": [...]
    },
    {
      "id": "watch-01",
      "type": "accessory",
      "color_names": ["gold"],
      "pattern": "solid"
    }
  ]
}
4 outfits generated in 128ms
10k/min
Rate Limit
<200ms
Response Time
2
Accessory slots per outfit

Powerful Features

Everything you need for intelligent outfit recommendations

AI-Powered Matching

Advanced style matching technology. Simply provide clothing images and get intelligent recommendations.

  • Auto-embedding generation
  • Lazy model loading
  • Accessory-aware scoring

Body Type Scoring

Intelligent fit recommendations for 9 different body types with scientific styling rules.

  • 9 body type profiles
  • Fit keyword matching
  • Silhouette analysis

Smart Filtering

Season-aware and context-based filtering ensures every outfit is appropriate.

  • Season matching
  • Occasion filtering
  • Keep-item preservation

High Performance

Optimized algorithms deliver lightning-fast recommendations at scale.

  • Sub-200ms responses
  • Efficient caching
  • Multi-worker support

Accessory-Aware Styling

Balance jewelry, bags, and finishing touches without overwhelming the look.

  • Up to two accessories per outfit
  • Respects keep-item requests
  • Duplicate-free combinations

Multilingual Support 🌍

Professional outfit explanations in 7 languages with culturally-appropriate fashion terminology.

  • English, Turkish, German
  • Spanish, Italian, Chinese
  • Azerbaijani

Confidence Ladder

Refined sigmoid mapping makes premium "very_compatible" looks attainable.

  • Clear label thresholds
  • LLM explanations optional
  • Metrics surfaced in responses

Ready to Get Started?

Join fashion platforms using our API to deliver personalized outfit recommendations.