ALIFF

Project Overview

A production backend for an AI modest-fashion stylist, transforming wardrobe photos into structured garment data and context-aware outfit recommendations through a multi-stage vision pipeline.

Engineering highlights

• •Five-stage computer-vision workflow for garment detection, classification, validation, and attribute extraction.
• •Multi-intent conversational engine routes styling requests through specialized AI pipelines from one endpoint.
• •Weather, wardrobe metadata, style preferences, and outfit history combine into adaptive recommendations.
• •Freemium quotas, token-cost tracking, and RevenueCat webhooks keep AI usage measurable and predictable.