SmartFit AR
A shoe marketplace where customers virtually try on shoes in AR before they buy — our BSIT capstone.
Overview
SmartFit AR is a web-based shoe marketplace that lets customers virtually try on shoes using augmented reality before buying. It connects local shoe shop owners with customers — complete with ordering, customization, feedback, and an admin control panel. This is the capstone thesis of Innovator Crews, BSIT 4th Year at Bataan Peninsula State University (AY 2025–2026).
The Problem
Buying shoes online is a gamble — shoppers can't try them on, so size and style mismatches drive returns, dissatisfaction, and lost sales. Local shoe retailers, meanwhile, have no unified digital storefront and rely on slow walk-in or phone orders, and there's no trust layer to verify which shops and listings are legitimate.
The Solution
SmartFit AR puts a real-time, in-browser AR try-on powered by DeepAR in front of a verified multi-shop marketplace. Customers browse shoes from approved local shops, try them on with their camera — no app install — and order in one place. Shop owners get a dashboard to list and fulfill orders, while an admin layer verifies every shop and listing before it goes live.
Key Features
- Real-time AR shoe try-on in the browser
- Marketplace connecting local shop owners and customers
- AI chatbot to guide and assist shoppers
- Ordering, customization, feedback, and an admin panel
Tech Stack & Why
- DeepAR API — Delivers production-quality AR try-on in the browser without building a CV pipeline from scratch.
- Firebase — Handles auth, data, and real-time updates across customers and shop admins with minimal backend code.
- AI Chatbot — Lowers the friction of a new shopping experience by answering questions inline.
- Vercel — Simple, fast hosting for the customer-facing app.
Biggest Challenge
Running accurate AR foot-tracking in the browser without a native app or a backend. We solved it by serving DeepAR .deepar effect files and shoe assets from Firebase Storage's CDN and loading them dynamically per shoe, so the AR engine streams only what it needs — keeping the experience install-free and light on student-grade devices.
Results & Impact
- Live capstone thesis, demonstrated to panel evaluators
- Led a 4-person team delivering all 12 planned features on schedule
- AR try-on across 3 user roles in a verified multi-shop marketplace
- Zero server cost on a fully static + Firebase architecture
What I Learned
Leading a capstone meant balancing ambition with scope — AR is impressive but only matters if the buying flow around it works. Coordinating a team, an advisor, and a real deadline taught me as much about project management as about code.
More Projects