Inspiration
E-commerce fashion loses billions due to checkout hesitation because static catalog photos cannot show how clothes fit. VibeIt eliminates this retail friction by turning text-based listings into an interactive, high-fidelity virtual dressing room.
What it does
VibeIt is a full-stack virtual dressing application split across three functional modules: • 3D Hero Page: An immersive landing layout powered by Three.js and high-performance GSAP scroll triggers. • Live Camera Studio: A custom webcam viewport that lets shoppers snap and process profile portraits instantly. • The Dressing Gateway: Users drop flat-lay clothes onto their photo to drape fabric configurations in 15 seconds. • History Gallery Vault: An authenticated page that securely logs past try-on runs for ongoing style analysis.
How we built it
• Frontend UI Layer: Built using Next.js 15 (App Router) and Tailwind CSS for responsive component scaling. • Core AI Gateway: Integrated the enterprise-grade Perfect Corp YouCam Clothes Virtual Try-On REST API (v3.0). • Unified Infrastructure Platform: Powered completely by InsForge via the native @insforge/sdk package to handle user Authentication, a PostgreSQL database ledger, cloud media Storage buckets (tryon-assets), and live global hosting. • Graphics Core: React Three Fiber (R3F) and GSAP handle real-time mesh distortions and stagger cascades.
Challenges we ran into
Converting binary live camera blobs into cloud URL links without slowing down the browser thread was a major hurdle. We also faced initial PostgreSQL Row-Level Security (RLS) policy errors when piping media links between YouCam endpoints and our storage bucket. We resolved this by defining clear folder access schemas natively inside the InsForge CLI configuration manager.
Accomplishments that we're proud of
• Zero-Lag Architecture: Engineered an automated workflow that handles fabric draping and database logs in 15 seconds. • Single-Platform Stack: Built a complete full-stack SaaS app without stitching together multiple separate cloud providers. • High-Fidelity Engagement: Kept website bundle load times light while running responsive, click-and-drag WebGL interactions.
What we learned
We mastered shifting heavy API scripts to server-side middleware paths to keep private keys hidden from client inspect tools. We also gained deep experience utilizing infrastructure-as-code CLI pipelines to verify, plan, and deploy cloud database networks instantly from the terminal.
What's next for VibeIt
We plan to scale VibeIt into an Agentic Omni-Stylist. By chaining YouCam’s Skin AI API alongside the current VTO tracks, an autonomous AI Agent will parse the saved camera photos, analyze the shopper's skin metrics and color season harmony (e.g., Cool Winter), browse a merchant's inventory catalog, and present personalized capsule wardrobe recommendations automatically.
