SAVR
A dining platform that recommends restaurants and dishes from your preferences, budget and company, and explains why each result fits.
- active users
- 10
- restaurants near Wolfville, NS
- ~50
- ways to ask
- 3
Historical figures as of 11 September 2026.
Problem
Restaurant apps filter by location and cuisine but ignore why you're going out, who you're with and what actually constrains the night.
Constraints
- Capture enough context without a heavy onboarding.
- Recommendations must be explainable, not a black box.
Decisions
Three ways to ask
Build Your Night (guided), Describe Your Night (plain language) and Surprise Me cover different moods without forcing one flow.
Every result states why it fits
Each match shows a fit score and the preferences behind it, so the user can correct the model instead of guessing.
Outcome
A live full-stack app with 10 active users: authentication, onboarding, saved presets, three recommendation flows and an AI-updated database of about 50 restaurants within 30 miles of Wolfville, NS. Relaunched in June 2026.
Evidence
- Live app and public source code.
- Screens captured from the running app.
Screens
Stack
FastAPI · SQLAlchemy · PostgreSQL · React · TypeScript · Vite · Playwright