Apify Actor 將美國 12 個城市及縣的建築許可資料整合至同一 API

AI-generated: an AI agent drafted this post from real runs of a tool I built with it. All the numbers come from those runs, for permits issued from 12 to 18 October 2026.

Building permits are some of the best leads there are. A new roof, a kitchen remodel or a 272-unit building is on record with the city weeks before the work starts. Most large US cities publish their permits as open data, but each portal has its own fields, date formats and quirks.

I built an Apify Actor that reads 12 of these portals into one table. Here is what one week looks like.

One week, 12 cities and counties

New York City shows "2,000+" because my run stopped at 2,000 rows. The counts are what each portal published for those dates when I ran the queries.

What stands out

Most permits are small trade jobs. Dallas issued 263 residential plumbing and 157 residential electrical permits that week. Austin issued 347 electrical and 283 plumbing permits. In Chicago, 416 of the 657 permits went through the express permit program. If you sell to plumbers, electricians or HVAC companies, this is your market, street by street.

Energy is everywhere. The busiest firm in Austin was Base Power, with 48 permits. In New York City, Momentum Solar came second with 21. In Orlando, Home Energy Group came first with 19. Solar, batteries and service upgrades show up in every city.

The big projects are easy to spot. The largest declared values of the week:

• New York City: $575 million for renovations to Arthur Ashe Stadium (122-02 Roosevelt Avenue);
• Austin: $245 million for a redevelopment filed under the city's SMART Housing program;
• Chicago: $71.65 million for an interior buildout of a 24-story building at 500 N Michigan Ave;
• Mesa: $44.8 million of tenant improvements in a 187,045 sq ft building;
• New Orleans: $21.2 million to stabilise a building envelope on Saint Peter Street.

The same row for every city

Each permit comes back in one schema, whatever the city calls its columns:

• permit number, type and status, plus a normalized status (approved, works completed, expired…);
• work description, address, ZIP code and coordinates;
• application, approval and issue dates;
• in metrics: the declared project value, the contractor's business name, housing units, the permit subtype and, in Austin and Seattle, a link to the city's permit page.

What each city publishes still differs. Dallas has no values or contractors to speak of, and Chicago, Orlando, New Orleans and Mesa have no ZIP codes. The Actor leaves those fields empty rather than guessing.

Two details I had to handle:

• Dallas publishes dates as text ("10/24/26"). A date range becomes a list of days in that format, so a Dallas query can span at most 400 days.
• Some cities mix people and firms in one field, for example "Jane Doe, Doe Construction LLC". Only the firm is kept. Names of individual owners and contacts are never collected.

Get it yourself

{
  "country": "US",
  "sourceIds": ["us-nyc-dob"],
  "zipCodes": ["10001", "10014"],
  "keywords": ["solar"],
  "maxResults": 100
}

The ZIP codes and keywords are filtered by the city portal itself, so a narrow query stays fast. Run one city at a time when you want a fixed number of rows per city.

From an AI agent, connect https://mcp.apify.com/?tools=gaelasao/planning-data-au-nz-fr and ask, for example: "List building permits issued in Chicago in the last 30 days for new construction, with the reported cost and the contractor."

It costs $5 per 1,000 permits. The Actor also covers France (permis de construire), NSW in Australia and Far North in New Zealand.

Which city should be next? Tell me in the comments.