MSU Hackathon 2026
A two-day hackathon idea for a breakfast shop, shown through a prototype that connects order review, receipts and ingredient planning.
Order OCR flowChatbot UIForecastingDashboard integration
A breakfast shop that runs on paper orders.
Our team of three built one system for it: a photo of a handwritten order becomes a structured order, a person checks the ones the AI is unsure of, then the orders add up into receipts, ingredient and stock forecasts, a chatbot, marketing help and a kitchen dashboard.
June was the qualifying round; the final was 13–14 August 2026.
One order, from photo to saved
The prototype, run locally with its demo data. The OCR step here is the built-in mock result, not a live reading of handwriting.
- 01
Read the order (mock OCR)
A sample pencil order slip goes in; the mock result comes back at 62% confidence, so the next step stays locked until a person checks it.
Prototype · demo data
- 02
A person checks it
After the check, the order total is worked out from the menu prices — ฿80 for this sample.
Prototype · demo data
- 03
Saved
The order is saved in the prototype’s session.
Prototype · demo data
- 04
Orders to review
The review screen with the order slip beside its details.
Prototype · demo data
Who did what
- Me
- In the final: the order OCR workflow and its data integration, the chatbot interface, the marketing assistant, ingredient and stock forecasting, the kitchen dashboard integration, and dashboard / expense fixes. In the qualifier: navigation, event filtering and pagination, animation, and the live-order screen.
- Team
- We decided the product, the OCR architecture and the human-review step together.
- AI
- Built very AI-heavy, on purpose, to move fast under the deadline. I used WhipUI during the event.
Winner · AI track · Higher Education.
A two-day prototype; the OCR in these captures is the mock path.