Inspiration

Across the United States, important information about a single property or worksite is scattered across federal agencies, state systems, local authorities, utilities, 811 services and public datasets.

The user should not need to understand that bureaucracy.

Our idea is simple:

You know the address. We find the rest.

UtilityDataUSA is built for anyone who needs reliable information about a U.S. address — homeowners, contractors, engineers, property professionals, utilities and businesses.

The goal is not to replace official authorities or field verification. The goal is to do as much of the information work as possible for the user, clearly show what is known, what is missing, and where authoritative confirmation is still required.

What it does

Enter a U.S. address and UtilityDataUSA builds a source-aware property and infrastructure profile from multiple public sources.

The existing platform already connects to sources including U.S. Census, FEMA, EPA, USGS, NWS and USDA, and provides guidance around 811 and other infrastructure-related information.

Instead of forcing the user to search multiple systems separately, UtilityDataUSA brings the evidence together in one understandable workflow.

Every result keeps its source, status and limitations visible.

A missing result is never presented as proof that no risk exists.

The challenge vision

For the Amazon Developer Hackathon, we are extending UtilityDataUSA from an address research platform into a more capable agent-driven experience.

The next layer is designed to understand not only the address, but also what the user is trying to accomplish.

A homeowner may be planning a swimming pool.

A contractor may be trenching fiber.

An engineer may be researching a site.

They all begin the same way:

One address. One search.

The agent can then use UtilityDataUSA's tools to investigate the relevant evidence, identify information gaps, preserve source limitations, and guide the user toward the next authoritative step when public data is not enough.

How it is built

UtilityDataUSA is built with Next.js and TypeScript and exposes both a human-facing web interface and agent-facing tools.

The existing application includes browser WebMCP tools and a remote MCP server using the MCP 2025-11-25 protocol.

The human interface and the agent tools use the same underlying source-aware data model, so the agent is working with the same evidence the user can inspect.

During this hackathon we are exploring Amazon's agent technologies to make the workflow more autonomous, contextual and useful while keeping the user in control.

What matters most

UtilityDataUSA is independent.

Public agencies, utilities and data providers are sources of information.

The user is who we are building for.

Our north star is to reduce the work required to understand an address while never pretending that public data knows more than it actually does.

The data may belong to many systems.

The answer should belong to the user.

Built With

  • epa
  • fema
  • mcp
  • next.js
  • nws
  • openai
  • react
  • supabase
  • typescript
  • u.s.-census-api
  • usda
  • usgs
  • vercel
  • webmcp
Share this project:

Updates

Submission history