How Find A Crib compiles its rent-regulation data
Every building on this site comes from a public record, and this page says which one, how it was joined to the rest, how often it changes, and what it does not prove. Four cities, four different registries, and they are not alike โ the differences are the most important thing on this page.
Find A Crib maps 131,289 records across New York City, San Francisco, Los Angeles and Washington, DC. New York and DC are official registries of rent-regulated buildings. San Francisco is an owner-reported inventory the city publishes anonymized to the block rather than the address. Los Angeles publishes no building-level list at all, so the Los Angeles layer is derived from assessor records against the city's own criteria and is labelled likely RSO. No listing here is a determination about a specific apartment.
What is on the map
| New York City | 47,165 buildings on the DHCR rent-stabilized register, 1,473,417 apartments by PLUTO's tax-lot unit counts |
| San Francisco | 12,297 block-side records from the Rent Board inventory |
| Los Angeles | 67,511 parcels meeting the RSO criteria (derived, likely RSO) |
| Washington, DC | 4,316 properties on the DC Rent Registry |
New York City โ DHCR register, joined to PLUTO and HPD
The starting record is the New York State Division of Housing and Community Renewal's 2024 rent-stabilized building file, which DHCR publishes as one PDF per borough. Those five files are parsed into rows keyed by Borough-Block-Lot (BBL), which is the join key for everything that follows.
- Coordinates, year built and apartment counts come from the NYC Department of City Planning's PLUTO extract. Unit counts are tax-lot figures, which is why every apartment count on this site is written as "about".
- Registered owner and managing agent, open housing-maintenance-code violations, and complaints come from NYC Open Data โ HPD's registrations (tesw-yqqr), registration contacts (feu5-w2e2), violations (wvxf-dwi5) and complaints (ygpa-z7cr) โ matched on BBL and linked back to the building's own record on HPD Online.
- Recently advertised units come from the overnight AffordableHousing.com feed, matched to DHCR-registered buildings. New York City only; there is no equivalent feed for the other three cities.
Data: NYC DHCR 2024 rent-stabilized building files, NYC PLUTO (coordinates), and NYC HPD Open Data (owner, violations, complaints). A building's rent-stabilized status reflects DHCR registration; it is not a guarantee that a specific unit is available or currently stabilized.
San Francisco โ reported to the Rent Board, published by the block
The source is the San Francisco Rent Board housing inventory, published by DataSF as dataset gdc7-dmcn: one row per unit per submission year, from the annual reports owners of units covered by the Rent Ordinance are required to file. Build years come from the San Francisco Assessor.
The anonymization is the source's, not ours, and it changes what a record means. DataSF publishes the address as a block ("1000 Block of Revere Ave") and the coordinate as the nearest mid-block point, so one San Francisco record is a block-side, not a building. That is why San Francisco is published here only as neighborhood aggregates: a per-address page would assert something the data does not contain.
Data: San Francisco Rent Board housing inventory โ units their owners reported as rent controlled, published anonymized to the block rather than the individual address, plus San Francisco Assessor build years. Coverage reflects what owners reported; it is not a determination that a specific unit is rent controlled.
Los Angeles โ derived, not published
Los Angeles publishes no building-level RSO inventory. The city's criteria, however, are mechanical: a residential property in the City of Los Angeles with two or more units, built on or before 1 October 1978. This site applies those criteria to the Los Angeles County Assessor parcel layer and keeps what matches โ 67,511 parcels.
Two limits worth stating plainly. Year built is tax-roll data, and the filter actually applied is assessor year-built of 1978 or earlier, which is marginally wider than an ordinance cut-off of 1 October 1978. And exemptions that exist in the ordinance โ luxury-exempt units, substantially renovated buildings โ cannot be derived from assessor records at all. Both are reasons this layer is labelled likely RSO and carries a verification link rather than an answer.
Data: Los Angeles County Assessor parcel records, filtered to the Rent Stabilization Ordinance criteria (two or more units, built before 1 October 1978). This is a derived list labelled "likely RSO" โ it is not the City of Los Angeles's official RSO inventory, which LAHD holds. Verify an address with LAHD or on ZIMAS before relying on it.
Washington, DC โ the Rent Registry, counted per unit
The source is the District's RentRegistry, run by the Rental Accommodations Division, which has published public export files of the whole registry since June 2025. Three files are used: one row per housing-provider registration, one per registered property, and one per unit.
Coverage under the Rental Housing Act is a per-unit determination, not a per-building one: a single property can hold both covered and exempt units โ subsidized housing, new construction, an owner of four units or fewer, co-ops and others. So coverage here is counted from the unit file, a property is kept when at least one of its units is not exempt, and the rent shown is the median registered rent of the non-exempt units. A registered rent is the legal rent on file with the District. It is not an asking rent and not what a unit is listed for.
Two derived fields to know about: DC publishes no ZIP on the accommodation record and no neighborhood polygons, so the ZIP comes from a Census ZCTA point-in-polygon test and the neighborhood is the nearest DC neighborhood label point. The neighborhood label is a nearest-point assignment, not an official boundary.
Data: District of Columbia Rental Accommodations Division registration and exemption filings (RentRegistry). A property appears here because its housing provider filed it as covered by rent control; registration is per property, and a unit its provider never registered is generally treated as covered rather than exempt.
How often each part changes
| Housing-voucher listings (/section8/) | Rebuilt nightly |
| HPD violations, complaints, owner and agent | Refreshed from NYC Open Data on the data build, not on every page build |
| DC Rent Registry | The source publishes nightly export files |
| SF Rent Board inventory | DataSF refreshes the extract every 24 hours; owners report annually, so the underlying filings move once a year |
| LA County Assessor parcels | The source layer refreshes monthly |
| DHCR rent-stabilized building files | Annual. The edition in use is 2024 |
Page-level dates follow the same rule. A page's sitemap lastmod is only moved when
that page's own content hash changes, so an unchanged building page keeps the date it last
genuinely changed.
What none of this decides
- None of it is legal advice, and none of it is a determination about a particular apartment. A building's presence on a register is about the building.
- Regulated status can change, and a registry reflects what was filed, at the time it was filed. The New York edition here is the 2024 file.
- Where a city's record is derived or anonymized โ Los Angeles and San Francisco โ the label on the page says so, and the verification route is named on the page rather than here.
If a record here is wrong, that is worth knowing and worth fixing: tell us what is wrong with it and name the address.
Using or citing this data
The New York corpus is downloadable in full as buildings.min.json, and there is a REST API for querying it. If you cite the figures on this site, cite the underlying registry as the source of the record and Find A Crib for the join and the derivation โ the agencies named above did the collecting.
Find A Crib was previously published as Jay's Home Finder (jayshomefinder.com) โ same rent-stabilization data, same maps, new name.