Live web app · React, Fastify, Postgres

Kalisi

The Bay Area food map that's already done the scrolling for you. Trending spots arrive pre-pinned and heat-ranked, with nothing to save or import before the map is useful.

Kalisi map of downtown San Francisco with dozens of heat-ranked restaurant pins
Every pin is a real spot creators are posting about. Color and size track its heat score.

Problem

Food recommendations were scattered across social posts, and existing apps asked for saves or friend invites before they became useful.

Decision

Resolve each video to a real place with an address and source links, then weight creator count, video count, views, and recency into one map score.

Result

More than 100 Bay Area spots on a live map, plus a weekly Heat Index that separates places gaining momentum from older viral hits.

Kalisi place card for Nan Hot Pot SF showing creator count, recency, and video links
A place card keeps the source videos, creator count, recency, and directions together.
Bay Area Heat Index ranking the fastest-rising food spots of the week
The weekly Heat Index ranks the fastest-rising spots across the Bay.

Under the hood

A React and Leaflet frontend runs on a Fastify API. Postgres stores places, mentions, and heat history. Ingestion, deduplication, scoring jobs, and the API run on Railway.

React / TypeScript / Leaflet / Fastify / Postgres / Railway / Vitest