Boston Curb Lab: using AI to turn confusing parking rules into a single public map

The problem

Every day, the curbs along Boston's streets are asked to do several jobs at once. They provide residents' parking, accommodate delivery trucks loading and unloading, and make room for accessible drop-off points, outdoor dining, bus stops, and bike lanes. As ride-hailing and online deliveries have grown, the competition for that limited space has intensified.

For a resident, a basic question - where can I park on this street and what are the rules- can be hard to answer. Boston's parking rules are intricate, vary from block to block, and have evolved over roughly 400 years. According to Michael Lawrence Evans, Director of the Office of Emerging Technology, “every delivery truck circling for a loading zone adds congestion and emissions. Every unclear sign and process creates unnecessary confusion and safety risks. These problems affect how people get around, how businesses operate, and the quality of life in our neighbourhoods.”

For city staff, the difficulty runs deeper. Boston's curb data has sat for decades in separate systems that do not connect, so planning any change - a new loading zone, a tighter parking restriction, a rerouted bus - has meant working across disconnected databases or sending someone out to survey the street. In practice, that has meant transportation planners checking Google Street View at their desks or walking a street, sign by sign, to match each one against engineering drawings.

City staff face a related problem. In order to plan and decide how curb space is used, city leaders, urban planners and researchers need dependable curb data to make sound decisions. But that data has been scattered across separate systems for decades, with no links between them, making it difficult to draw on a single, reliable view of the city's curbs.

The consequences reach beyond inconvenience. Sam Brenner, a Transportation Technology Strategist in the office, notes that poor experiences at the curb erode residents' trust in the city, with some coming to feel the authorities are looking to ticket or tow rather than help. Clearer curb information, the team argues, is a chance to reverse that impression.

The solution

In February 2026, the City of Boston launched the Curb Lab, an initiative that uses AI and open data to build something the city has never had: a single, continuously updated record of the rules governing every curb in Boston. The work is funded by the U.S. Department of Transportation SMART Grant awarded in 2024.

The launch followed more than a year of preparation. The city first worked with the UrbanismNext research team at the University of Oregon to determine which curb data mattered most and where it could do the most good. It then built and tested prototypes with residents and staff, among them apps that help people make sense of the city's most confusing parking signs. One of those prototypes, an early version of the public parking map, helped the team decide what to build first. Putting it in front of users showed that no map or app would be worth using unless the curb data beneath it was accurate and complete. The office therefore chose to validate the underlying data before refining how the tools looked and worked for the public. In partnership with the consultancy Cambridge Systematics, it built the underlying data systems. Pilots are now running across four areas: curbside safety, resident accessibility, city efficiency and local economic development.

At the core is computer vision, a form of AI that recognises features in images. The system reads street-level photographs of Boston's roads, identifies parking signs, road markings, and other features, and converts them into structured digital data. That record is refreshed each night, so it stays close to current. For residents, the Curb Lab is building a public map of the parking rules for every street. Delivery drivers will be able to locate legal loading bays, and people with disabilities will be able to find accessible drop-off points.

The entire system is built on the Curb Data Specification, an open standard from the Open Mobility Foundation that gives cities and companies a shared language for curb data. Because of it, Boston's data can flow straight into the navigation, ride-hailing and delivery apps people already use. According to the Open Mobility Foundation, Boston is the first city to build and deploy curb-management tools at this scale using the specification, and the city has committed to releasing all its code publicly so others can take the same approach without having to rebuild it themselves.

The city has also worked with CodeMetal on inexpensive sensors that gather curb and traffic data in real time without identifying the people on the street. As the project is still young, the team is building the measures it will use to judge success, such as how heavily curb space is used, how often drivers circle for a space, and how often delivery vehicles double-park or block a lane.

The impact

1. Centuries of parking rules are being converted into a searchable digital map for the first time
Boston's parking regulations have accumulated over more than 400 years. The Curb Lab is converting them into a single, publicly accessible digital map. Once it is live, a resident will be able to check the rules for a given street directly, instead of piecing them together from the signs on the block, and the same map will show delivery drivers where they can legally load and people with disabilities where the accessible drop-off points are.

2. Disconnected city systems are being joined up.
Curb data that has spent more than 40 years in separate, unconnected systems is being brought into a single view. In the words of Amelia Capone, the city's Director of Parking and Curbside Management, linking them means “less duplication of work, faster response times to constituent requests, and better internal coordination.” Its value to city staff comes from joining up information that used to sit in separate systems. Roadworks and temporary closures, transit diversions, construction permits, outdoor seating areas, food-truck pitches, sensor readings and parking-payment records now appear in one place, so staff can see what is happening on the curbs without having to move between systems.

As the Curb Lab launched in February 2026, these are changes now underway rather than measured results. Further updates on these initiatives as they progress will be available on the Curb Lab webpage in the coming months.

Lessons learned

• The problem was not a lack of rules but a lack of visibility. Boston's curb regulations were already detailed and extensive. They had simply never been gathered into one place a person could see, because they lived on physical signs and in systems that did not connect. Digitising them leaves the rules themselves untouched while making them visible to residents and usable by staff for the first time. For other governments, much of the value in this kind of work lies in making existing rules legible.
• Time spent on groundwork made the build possible. Boston did not begin by building. Its first phases, from late 2024 into the summer of 2025, focused on determining which curb data was needed and on testing prototypes with residents and staff. Only after that did it build the data systems, through late 2025, and open its pilots in early 2026. For cities considering similar work, the early stages of research and user testing are what make the later, more ambitious applications work.
• Designing for reuse from the start lowers the cost for everyone else. By building on the Curb Data Specification and committing to publishing its code, Boston has set up its system so that other cities can follow without having to build everything themselves. The Open Mobility Foundation's SMART Curb Collaborative provides cities with a forum to learn from one another as they tackle the same problems.
• Reliable curb data is useful well beyond parking. Because so many modes of travel meet at the curb, a dependable record of it can inform far more than parking. The same data feeds into delivery logistics, city efficiency, accessibility, and congestion, making it a foundation for more joined-up decisions about how a city's streets work.
• Accurate data depends on people, not just technology. The team's clearest lesson for others is that maintaining reliable curb data rests on the staff who change signs on the ground and on teams sharing information through a common system. Keeping the record accurate is a matter of relationships and coordination between groups that may not have worked together before, rather than of the technology alone.

This case study was written with assistance from artificial intelligence.