For researchers & builders

License the dataset

Every coverage figure on this page was counted directly from the production database on July 31, 2026. The enrichment pipeline runs nightly, so counts move upward between recounts.

Churches List maintains a heavily cross-referenced database of American Christian congregations: 334,555 places of worship, each joined where possible to public registries — the Census, the National Register, the Library of Congress, the IRS Exempt Organizations file, Wikidata and OpenStreetMap. The same data that powers this directory is available as flat-file exports and as an API.

Take the sample first

No form required to look at the goods:

  • 500-row sample, CSV — a deterministic cross-section of the full file, every column populated exactly as it ships.
  • Data dictionary, CSV — every field with its type, source, row count, coverage percentage, and the licence of the underlying source.

Coverage, counted honestly

Coverage is uneven and we would rather you know that before you pay. The core address and geography layer is near-complete; the historic and financial layers are deep but narrow, because a precision-first match that leaves a field empty is more useful than a loose one that fills it wrongly.

LayerRecordsCoverage
Name, city, state, denomination334,540–334,555100%
ZIP code334,00699.8%
Latitude / longitude330,41198.8%
Street address327,99998.0%
County name + FIPS327,26897.8%
County-level US Religion Census context327,26597.8%
Census ACS profile for the ZIP327,26897.8%
Metro area (CBSA)226,15367.6%
Wikimedia Commons media links84,86425.4%
GNIS feature match82,40324.6%
Internet Archive first-snapshot date71,53221.4%
IRS Exempt Organizations match60,72318.2%
Nearby-POI profile (OpenStreetMap)39,64711.9%
Website URL97,33029.1%
Chronicling America newspaper mentions15,6394.7%
Google-verified profile corroboration11,3633.4%
Wikidata founding year2,4130.7%
Wikidata architectural style2,3270.7%
National Register of Historic Places2,1750.7%
Wikidata architect attribution1,2820.4%
Historic American Buildings Survey3980.1%

Two figures worth reading twice. The National Register layer is 2,175 congregations, not the ~9,500 an address-only match produces: we replaced the loose match with a distance-and-name-similarity match in June 2026 and threw the difference away. And of the 60,723 IRS matches, only 211 currently carry a Form 990 financial filing — most congregations are exempt from filing one. If your product depends on church financials, this is not the dataset for it.

What is deliberately not in it

No personal data. We do not license member lists, donor information, clergy contact details, or the correspondence churches send us. Exports carry organizational records only — the same information a visitor sees on a public church page, plus the public-registry joins.

Formats

  • CSV or Parquet — full file or a filtered slice (by state, denomination, or enrichment layer), with the data dictionary and per-field source licences.
  • API — JSON over HTTPS, keyed, for products that need fresh reads rather than a snapshot.
  • Custom cuts — a specific join, say National Register churches with a website in one metro. Most cuts take us a day.

Licensing

Academic & newsroomCommercial snapshotLive access
From$99 one-time$499 one-timePriced by volume
Full core fileYesYesYes
All enrichment layersYesYesYes
Data dictionary + provenanceYesYesYes
Use inside a commercial productNoYesYes
API accessNoNoYes
RefreshOne snapshotOne snapshotChange feed
Redistribution of the raw fileNoNoNo
AttributionCitation requiredRequestedRequested

Students and small newsrooms: write to us anyway, we are flexible. Building retrieval or agentic products that need authoritative US church data? Ask about bulk snapshots plus a change feed — see also our llms.txt.

The underlying public sources (IRS, Census, NRHP, HABS, Chronicling America, OpenStreetMap, Wikidata) carry their own terms, and the data dictionary marks the provenance and licence of every field. What you license from us is the compilation, the matching, and the corrections layer on top.

Why this is hard to reproduce

The list of names is the easy part. The value is in the joins: tying a congregation's street address to its IRS record, its National Register entry, and its 1890s newspaper mentions is slow, error-prone work that the pipeline re-verifies nightly — plus a human corrections loop, where churches themselves submit fixes that we verify and apply, usually within a day or two.

Request a quote

Tell us the use case and the slice you need. We answer with a quote and, if it helps, a larger targeted sample.







Prefer email? letters@churcheslist.com, subject “Dataset licensing”.

A note on neutrality: licensing data does not buy placement, ranking, or editorial treatment in the directory. The dataset and the editorial product are separate; our principles apply to both.