Vishal Deshmukh Information Systems Grad student vishald089@gmail.com
Campus Shareable Kits — Hub-first Aggregator
Borrow smarter, waste less. The app helps students find the best way to borrow (official library hubs first, peer fallback if needed) and shows the sustainability impact of choosing not to buy.
The problem (why this matters)
On most campuses, libraries and makerspaces lend gear—but hours, stock, and scope are limited. The moment a hub is closed or out of stock, people give up and buy a charger, projector, or toolkit they’ll use once. That creates waste, duplicated purchases, and extra cost. The “green” option isn’t the easiest one; it’s fragmented across websites, policies, and availability pages. Students need a single place that routes them to the best lending option right now—and makes the sustainable choice obvious.
What the app does (what it feels like to use)
Open the app and you’ll see a short Intro card (“Borrow smarter, waste less.”) and a clear Discover view. Search “projector,” for example:
The app’s BestSource logic checks official hub items and hours. If a library item is available and the hub is open, you get a big “Go to Hub (Policy)” button with a plain-English policy summary and the official link.
If the hub is closed or out of stock, you’ll immediately see peer kits nearby and a “Request to Borrow” flow (with request → approve/decline → collected → returned → review).
An Impact Dashboard tallies CO₂e avoided, items borrowed, and money saved using simple, transparent factors (e.g., 50 kg CO₂e for large AV gear).
First-time users get a light tutorial overlay (with a working “Got it”/“Skip” control) that points out the four key areas: Discover, Item Details, My Requests, and Impact.
Why AI (and how we keep it trustworthy)
This isn’t a black box. The app uses deterministic, agent-style flows that are transparent and predictable:
BestSource (hub vs peer) prefers official hubs when hours + inventory match your window; otherwise it falls back to peer items.
PolicyExplainer turns a policy page into a short, friendly summary (with the source link).
ImpactEstimator uses published, fixed factors (e.g., chargers = 2 kg, tools = 8 kg, AV = 50 kg, party = 5 kg, other = 3 kg CO₂e) and a simple avoided-purchase assumption to compute impact.
Borrow-vs-Buy gives a nudge based on expected use frequency.
That combination makes the app feel “smart” without hallucinations or surprise behavior.
Design choices (why it’s easy to use)
I aimed for a calm, modern look that makes decisions obvious:
Glassmorphism surfaces with soft blur and teal/green accents (sustainability cue).
Clear badges: blue = Official hub, green = Peer. Status pills: available (green), reserved (orange), out (gray).
Three-tap flows: request → approve → returned, with crisp toasts and micro-animations.
Accessibility: large touch targets, logical focus order, and friendly copy everywhere.
Technical execution (how it’s built)
Built entirely in Base44 with collections for users, orgs, org_items, kits, requests, reviews, and impact_logs. Server actions handle the request lifecycle; AI flows are small and deterministic.
About login (intentionally handled for demo): For the hack demo, browsing is public so judges can open the app and see items immediately. Write actions (creating a request, approving, marking returned, reviewing) require sign-in. This is intentional: it lowers friction for judging while still demonstrating permissions.
How login works when we scale: In production we’ll enforce campus SSO (e.g., SAML/OAuth) and strict Row-Level Security:
Roles: Student (borrower), Peer Owner, Org Manager (for library/makerspace items), and Admin.
RLS:
Students can read all public inventory but only modify their own requests.
Peer Owners can edit their kits and approve/decline requests to them.
Org Managers can manage org_items for their hub and see analytics for their org.
Audits & trust: reputation scores from reviews, optional deposits/holds for high-value items, and late-return safeguards.
Integrations: campus directory (to verify enrollment), hub availability feeds, and optional reminder emails/calendar holds.
This path preserves today’s smooth demo while making the production system secure and enterprise-ready.
Expected impact (measurable benefits)
CO₂e reduction: If 100 students borrow small items instead of buying, that avoids roughly 200–800 kg CO₂e per term. With larger AV gear, a single borrow can avoid ~50 kg CO₂e.
Money saved: Replacing one-off purchases saves $20–$150+ per item; AV gear and tools save more.
Time saved: One search replaces hopping across multiple sites and policy pages; fewer dead-end trips to closed hubs.
Utilization: Official inventory is used more consistently, and peer kits fill night/weekend gaps.
Scalability (beyond one campus)
The model generalizes to any community with lending hubs—other universities, coworking spaces, apartment complexes, or city tool libraries. Hubs are just “orgs;” their items are “org_items.” The peer layer fills gaps automatically, and the same BestSource logic scales to many hubs across a city.
Impact: Makes the sustainable option the default, not the “extra effort.” Shows real numbers for CO₂e and dollars.
Innovation: Aggregates official hubs and peers, adds routing + policy summaries + transparent impact math.
User Experience: Zero-friction intro, optional tutorial, clear badges/pills, and three-tap request flows.
Technical Execution: Clean Base44 schema, guarded writes, deterministic AI flows, and self-healing seed logic for reliability.
Log in or sign up for Devpost to join the conversation.