Alta Daily, an AI-powered fashion app, has processed more than 20 million images using Meta's Segment Anything Model (SAM) to segment and digitize users' clothing (Meta AI Blog). The app allows users to photograph their wardrobes and receive outfit recommendations based on natural language prompts, displayed on a personal digital avatar. Most people wear only an estimated 20% of the clothes in their closets (Meta AI Blog), and Alta Daily was built to unlock that untapped potential by making it easier to visualize and remember outfit combinations.
Alta Daily founder and CEO Jenny Wang chose SAM after testing various segmentation models across eight product categories, finding that SAM consistently delivered the best results on challenging real-world scenarios such as white sneakers on white walls and reflective surfaces (Meta AI Blog). The decision to use SAM had significant financial implications: external segmentation APIs cost a few cents per image (Meta AI Blog), which would accumulate rapidly at scale. By adopting Meta's open-source model, Alta Daily reduced infrastructure costs while maintaining the clean, editorial-style interface that defines the user experience.
For commerce and fashion-tech practitioners, this case illustrates how open-source foundation models can democratize access to sophisticated computer vision capabilities, enabling startups to compete on product quality rather than API budgets. The app has already gained a global following across the United States, France, Germany, Mexico, and the Netherlands (Meta AI Blog), suggesting strong product-market fit in the digital styling space.