About RealEstateDataset

We compile publicly available U.S. property records into a single, standardized dataset covering roughly 146 million properties. You can download the whole country, or slice it down to a single state, county or ZIP code — as Parquet (with full history) or CSV (flat).

Every file is released under CC0 1.0: public domain, no attribution required, free for commercial use. No sign-up, no gate, no catch.

Property data has always been public in principle and locked up in practice — scattered across thousands of county systems, or bottled behind expensive enterprise contracts. We think the baseline facts about American homes should be a public good. So the bulk dataset is free, forever.

We cover the facts about each property — address, beds, baths, square footage, year built, last sold price, listing status. We do not publish owner names or any personal information.

A downloaded file is a snapshot — accurate the day it was compiled, and slowly going stale after that. If you need current data, or you'd rather query one property at a time than host 100 GB, that lives on Straply — the real estate data API this dataset is built from. Free tier included; the same records, always fresh.