← Back View Code

AisleIQ — Retail Price Intelligence

Catapult 2026 Hackathon

Think a cross between Honey (the online rewards extension) and Klaviyo (the eCommerce customer-behavior app) — but for physical retailers. Grocery pricing increasingly changes day to day based on demand, so shoppers need an edge; meanwhile in-store brands lack visibility into how customer behavior translates to purchases or missed opportunities. AisleIQ bridges that gap with real, user-contributed data.

Vendor analytics overview dashboard
Vendor analytics overview
Pricing intelligence dashboard
Pricing intelligence
Consumer product-scan and price comparison screen
Consumer scan & price comparison

What it does

A dual-platform system:

  • Consumer app — scan products, compare prices between stores, and predict when and where your cheapest grocery trip will be.
  • Vendor platform — AI-powered analytics, competitive insights, and ML-driven pricing optimization tuned to goals like market share or revenue.

Nothing on the market delivers this level of human-to-product data; the closest analog is DoorDash recently asking dashers to photograph shelves — a signal of a market just opening up.

How we built it

Two Next.js apps connected via Firebase for real-time data, plus a FastAPI service running a TinyTimeMixer model we trained for price forecasting and optimization.

Built with

Next.js React TypeScript Node.js Firebase FastAPI PyTorch

Challenges

  • No reliable real-time retail pricing APIs.
  • Designing business logic that gets valuable data while benefiting the contributor.
  • Integrating ML predictions into a live dashboard.

Proud of

  • End-to-end pipeline from scan → insight.
  • Real-time dashboard with meaningful metrics: which day to plan your next trip (shoppers) and what price/placement wins market share (vendors).
  • Working ML-based price prediction and weekly simulation.

What we learned

Real-world data is messy, but intent is what's valuable: knowing a customer walked in, debated products X and Y, and left with Y tells a richer story than stock counts — which is all the industry currently provides. Building something where users gain from the data they contribute (rewards on every receipt, product, and shelf scan → a cheaper haul) is a far friendlier way to collect it.

What's next

Collect daily data from surge-pricing grocery stores, improve data quality and user adoption, partner with retailers on discounts and rewards, and explore how this data could feed robotics.