Style
The Style domain gives your AI assistant a personal stylist's understanding of your wardrobe. Through Fluent, your assistant can work from every piece you own, your aesthetic preferences, how you dress, and where the gaps are — so purchase decisions, outfit choices, and wardrobe advice are grounded in what you actually have.
What Style Knows
- Your closet: Every item with category, brand, color, size, status, and comparator key
- Fit profiles: Typed descriptors for silhouette, visual weight, structure level, texture, fabric hand, polish level, and quality tier
- Photos: Media with typed roles (product shot, fit photo, detail) delivered through MCP
- Style profile: Your preferences, body context, and calibration state
- Provenance: Field evidence, source snapshots, and technical metadata for each item
Key Workflows
Closet Management
Your assistant can list, add, update, and inspect items in your closet. Each item has a canonical comparator key that Fluent uses for deduplication and overlap detection.
"List everything in my closet"
"Add this new shirt I just bought"
"What do I have in outerwear?"Purchase Analysis
When you are considering a purchase, Fluent provides coverage-aware wardrobe context. The assistant inspects the candidate image alongside relevant closet item images before making a recommendation.
"Should I buy this jacket?" (with image)
"How does this compare to what I already own?"The style_analyze_purchase tool provides comparator-coverage-aware context, and style_get_visual_bundle delivers the images for visual grounding. The assistant then makes the recommendation with an actual stylist point of view.
Wardrobe Analysis
Fluent can analyze your entire wardrobe for gaps, redundancies, bridge pieces, and buy-next guidance.
"What are the weak spots in my wardrobe?"
"What should I buy next?"
"Do I have too many black t-shirts?"Descriptor Enrichment
Fluent tracks which items need visual descriptor enrichment and prioritizes them. Fit photos drive fit observations, silhouette, visual weight, and structure level. Product photos drive texture, fabric hand, polish level, and quality tier.
Style Tools
| Tool | Purpose |
|---|---|
style_get_profile | Read your style profile and calibration state |
style_update_profile | Update style preferences and body context |
style_get_context | Full style context for the assistant |
style_list_items | List closet items with filters |
style_get_item | Get a single closet item |
style_get_item_profile | Typed fit profile for an item |
style_get_item_provenance | Field evidence and source metadata |
style_upsert_item | Add or update a closet item |
style_upsert_item_profile | Write typed descriptors for an item |
style_upsert_item_photos | Attach photos with typed media roles |
style_analyze_purchase | Coverage-aware purchase analysis |
style_get_visual_bundle | Retrieve images for visual grounding |
style_analyze_wardrobe | Gap lanes, replacements, buy-next guidance |
style_list_evidence_gaps | Items missing photos or descriptors |
style_list_descriptor_backlog | Prioritized enrichment queue |
Style Resources
| Resource | Purpose |
|---|---|
fluent://style/profile | Style profile and preferences |
fluent://style/context | Full style context |
fluent://style/items | All closet items |
fluent://style/items/{item_id} | Single item detail |
fluent://style/item-profiles/{item_id} | Typed fit profile |
fluent://style/item-provenance/{item_id} | Provenance and evidence |
Onboarding
Style onboarding is calibration-aware:
- Seeded users (imported closet): Onboarding is calibration over existing items
- Fresh users with a purchase candidate: Start with purchase analysis
- Fresh users without items: Collect a small anchor set of items first
After onboarding, the domain moves to ready and the full surface is available.