Permit Leads Across 4 Cities, One Shared Format

Proptech and multi-market contractor-supply analysts get building permits from four major-city portals under one shared set of column names, ready to benchmark without writing four separate parsers.

Introduction

Pick one city and a date range. Get back one row per building permit the city issued in that range, with the link to the city's own public record on every single row.

Four cities are supported, and only four: Chicago, Los Angeles, Boston and Washington DC. This is not a nationwide permit feed. If a project needs Houston or Phoenix, this Actor is the wrong tool.

Each run reads the city's own open-data portal live, at the moment the run starts. Nothing is served from a stored copy. The four portals behind it are the City of Chicago data portal, the City of Los Angeles data portal, Analyze Boston and Open Data DC. All four are free and public. What this Actor adds is one shared set of column names across four portals that otherwise disagree about almost everything, a date filter that works the same way in all four, and a record link on every row so any number can be checked at the source.

Every row carries these fields:

Field	What it holds
city	Which of the four cities the row came from
permit_number	The city's own permit number
permit_type	The city's permit type wording, unchanged
permit_subtype	Review type, occupancy type or permit subtype, where the city publishes one
work_class	Work type or category, where the city publishes one
status	The city's status wording, where the city publishes one
issue_date	Date the permit was issued, as YYYY-MM-DD
address	Street address as the city published it
zip	Postal code, where the city publishes one
ward	Ward or council district, where the city publishes one
valuation_usd	Declared or reported job value in US dollars, as a number
fee_paid_usd	Permit fees in US dollars, as a number
square_feet	Floor area, where the city publishes one
work_description	The city's description of the work
latitude, longitude	Point location, where the city publishes one
attribution	The credit or notice the publishing city requires
source_dataset	The portal and dataset the row was read from
source_url	Link to that exact record on the city's public API
fetched_at	UTC time the row was read, as an ISO timestamp

A field the city does not publish arrives as null. It is never guessed and never filled in from somewhere else.

Tutorial

1. Choose City from the dropdown.
2. Set Issued on or after and Issued on or before. Both days are included in the results.
3. Set Maximum rows. This is a hard ceiling on rows returned, so it is also a hard ceiling on what the run can cost. It starts at 100, so a run left on the default costs at most $0.30.
4. Start the run.
5. Open the Permits output to browse rows, or Download permits as CSV to take them away. The Run status output holds the row count, the oldest and newest issue date actually returned, the source, the required credit line, and any warnings.

To sweep a long period without one enormous run, run month by month and raise Maximum rows only as far as needed.

Pricing

$0.003 per row, charged from the first row. There is no start fee.

Billing is per result. One row pushed is one charged result. A run that returns 40 rows is charged for 40 rows, which is $0.12. A run that returns 100 rows costs $0.30. A run that returns 1,000 rows costs $3.00.

Three kinds of run push no rows, and a run that pushes no rows costs nothing:

• A run that fails because a city portal is down. Status is UNKNOWN, the exact error is reported, and no rows are pushed.
• A run rejected for bad input. Status is INVALID_INPUT and no rows are pushed.
• A run where the city genuinely issued no permits in that date range. Zero rows pushed is zero charged.

Because nothing is charged except a pushed row, Maximum rows is a real spending cap. Set it to 100 and a run cannot cost more than $0.30.

Apify handles billing and account charges. Apify platform usage is separate from the per-row price.

Input and output examples

Input, which is also example.json:

{
  "city": "chicago",
  "issuedAfter": "2026-08-01",
  "issuedBefore": "2026-08-31",
  "maxItems": 100
}

One output row, trimmed for length. The full three-row sample is in sample-output.json:

{
  "city": "chicago",
  "permit_number": "101076242",
  "permit_type": "PERMIT - RENOVATION/ALTERATION",
  "permit_subtype": "STANDARD PLAN REVIEW",
  "status": "ACTIVE",
  "issue_date": "2026-08-31",
  "address": "636 N ST LOUIS AVE",
  "zip": null,
  "ward": "27",
  "valuation_usd": 114880.0,
  "fee_paid_usd": 3099.11,
  "work_description": "SPR 2019 CBRC: INTERIOR AND EXTERIOR RENOVATIONS ...",
  "latitude": 41.892969033692374,
  "longitude": -87.71394557854016,
  "source_dataset": "data.cityofchicago.org/resource/ydr8-5enu",
  "source_url": "https://data.cityofchicago.org/resource/ydr8-5enu.json?id=3431525",
  "fetched_at": "2026-09-15T02:31:02Z"
}

Opening that source_url returns that one permit from the City of Chicago's own API. Every row in every city works the same way.

Limitations

Read this part before paying for anything.

• Four cities, not the United States. Chicago, Los Angeles, Boston and Washington DC. Nothing else.
• Each city's history starts where its own portal starts. On 14 September 2026 the portals held issue dates from: Chicago 3 January 2006, Los Angeles 1 January 2020, Boston 26 September 2006, and Washington DC calendar year 2009 onward. Asking for dates before a city's start returns fewer rows, not an error.
• Freshness is the city's, not this Actor's. Portals lag by days, and each city lags differently. On 14 September 2026 the newest issue date held was 13 September for Chicago, 12 September for Los Angeles and 3 September for Boston. A permit issued yesterday may not be on the portal yet. fetched_at records when the portal was read, not when the city last updated it.
• No names of people. Applicant, owner, contractor and contact-name columns are removed before rows are built, in every city that publishes them. This Actor returns permits, not people, and it is not a contact list.
• Washington DC has no work description. The District puts phone numbers, email addresses and contact names inside that one free-text field, so the whole field is dropped. work_description is always null for Washington DC.
• Fields differ between cities because the cities differ. Chicago publishes no postal code on this dataset. Washington DC publishes fees but no job valuation, and no work description. Those arrive as null, not as a substitute value.
• maxItems cuts the newest rows first. Results come back newest issue date first. Hitting the ceiling means older permits in the range were not returned, and the run status says so in its warnings.
• These records are the city's, and cities revise them. A permit can be amended, voided or reissued after it appears. This reports what the portal held at fetched_at, and makes no claim the city's own table is complete or correct.
• This is not a licence to the underlying data. The charge is for the collection and the shared schema. Each city's own terms still apply to what is done with its rows, and each portal remains free to use directly.

Recommendations

Start with one month in one city at the default 100-row ceiling. That run costs at most $0.30, and the rows show exactly what that city does and does not publish before more is spent.

Filter on permit_type after the fact rather than expecting the same wording in two cities. The type wording is each city's own and it does not match across portals: Chicago writes PERMIT - NEW CONSTRUCTION, Los Angeles writes Bldg-New, and Boston writes Short Form Bldg Permit.

Treat valuation_usd as the value declared to the city, which is what it is. It is not an appraisal and it is not a construction cost estimate.

Keep the attribution field with the rows. Chicago and Washington DC both require a notice to travel with any copy of their data, and that field carries the exact wording each one requires.

For a repeating pull, schedule one run per city per week over the last seven days, rather than one large backfill run.

FAQ

Does this cover my city? Only if it is Chicago, Los Angeles, Boston or Washington DC.

Can I get contractor or owner contact details? No. Person-name and contact columns are stripped before rows are built, in every city. This Actor is not a contact database.

Is the data live? Each run reads the city portal at the moment it starts. How current that portal is remains the city's decision, and the Limitations section gives the lag seen on 14 September 2026.

What happens if a city portal is down? The run reports status UNKNOWN with the exact error from that portal, pushes no rows and charges nothing. Nothing is invented and nothing stale is substituted.

Why is a field empty? Because that city does not publish it on that dataset, or left it blank on that record. Empty fields are never filled in with a guess.

Can I verify a row? Yes. Open its source_url. It returns that one record from the city's own public API.

Can I get this data for free? Yes. All four portals are free and public. Going direct means four different APIs, four sets of column names, four date-filter syntaxes and no shared schema. That work is what the per-row price covers.

Is there support? Use this Actor's Issues tab with the run ID and a redacted example. There is no promise of a reply from a person and no service-level agreement.

Source credit

• Chicago: City of Chicago, dataset ydr8-5enu. The City's required data notice travels on every Chicago row in the attribution field.
• Los Angeles: City of Los Angeles, dataset pi9x-tg5x, Building Permits Issued from 2020 to Present.
• Boston: City of Boston, Analyze Boston approved building permits, under the Open Data Commons Public Domain Dedication and License (PDDL) v1.0.
• Washington DC: Open Data DC, Building Permits, under Creative Commons Attribution 4.0 International (CC BY 4.0). The required credit travels on every Washington DC row in the attribution field.

Refunds and support

Refunds: if a run returns zero rows for a valid input, email operations@ustechautomations.com within 7 days and we refund that run. Support: same address, replies within 2 business days.