About
Computer vision. City infrastructure. A better curb.
Parkspot.ai turns infrastructure that already exists into real-time data. We think cities, fleets, and drivers deserve to know what's actually happening on every block — not a guess, not a five-minute-old average, and not a third-hand survey.
Looking for our mobile app? DriversMobile app·Cities·Fleets
Our story
Founded by people who argue about curbs at dinner.
It started with a simple observation: in busy parts of the city, a curb spot can take thirty minutes to find. Drivers idle. Cars circle the block. People end up double-parked, triple-parked, on the sidewalk. It's inefficient, it's bad for the street, and it's bad for the air. Our bet: government-operated cameras in NYC already carried the signal. The same block-face timeline now powers a driver app, a city operations tool, and a fleet dispatch layer.
Sustainability
Three levers, measurable.
When parking gets predictable, three things drop: fuel consumption, CO2 emissions, and noise. SFMTA's audited SFpark pilot (2014 evaluation) measured ~2 tonnes of CO2 avoided per day across a ~0.5 sq mi zone through demand-responsive pricing and availability info. A guidance-first deployment in a Midtown-core study area projects ~1–2 tonnes of CO2 per day under the same mechanism.
Principles
We read feeds. We don't store them.
We read camera streams from two sources: government-operated cameras the city already shares, and opt-in feeds from individuals or businesses who choose to contribute their own. Every feed is ingested with the operator's explicit consent. We don't store video, capture plates, or recognize faces.
Our vision
The curb is infrastructure. Treat it that way.
A parking bay is as much infrastructure as a lane of traffic or a bike lane. Cities that see curb data in real time can reprice, rezone, and reclaim space with evidence. Fleets that see availability can route around the parking hunt at every stop. Drivers stop circling — a behavior that averages 7.5 minutes per driver across 21 downtown studies.