IRS Nonprofit 990 Scraper — Charities, Revenue & Officer Pay

Turn the US nonprofit sector into a spreadsheet. Search official IRS Form 990 data by keyword, state, cause area and 501(c) type, and get one clean row per organization: name, EIN, address, cause area, 501(c) type, and — straight from the organization's own filings — total revenue, expenses, end-of-year assets, surplus/deficit, operating margin, officer compensation and year-over-year revenue growth, plus a link to the source 990 PDF.

No API key, no login, no proxy. Public IRS data via ProPublica's free Nonprofit Explorer API.

What you can do with it

• 💰 Qualify prospects by size — minRevenue + the size_tier column ("mid ($1M-10M)") give you only the nonprofits big enough to be worth a pitch.
• 🎯 Build a targeted list — every food bank in Texas, every health charity in Wyoming, every 501(c)(6) chamber of commerce in your state.
• 📈 Spot who is growing — revenue_growth_pct compares the latest filing to the prior year, so you can find organizations whose budgets are expanding.
• 🔎 Check financial health — surplus and operating_margin_pct show at a glance who runs a deficit and who does not.
• 🧾 Research officer pay — officer_compensation comes from the filing itself.
• 🗂️ Export to CSV, JSON or Excel.

Built for fundraising consultants, grant writers, nonprofit-sector vendors, researchers and journalists who need the whole segment rather than one EIN lookup at a time.

Input

Field	Type	Description
searchTerms	array	Keywords matched against organization names, e.g. food bank, animal rescue. Leave empty to browse a whole state or cause area.
states	array	Two-letter US state codes, e.g. ["TX","CA"]. Empty = all states.
causeArea	select	One NTEE major group — Arts, Education, Environment & Animals, Health, Human Services, International, Public Benefit, Religion, Mutual Benefit, Unclassified.
subsectionCode	select	One 501(c) type. 3 is the classic tax-deductible public charity.
includeFinancials	boolean	Fetch each organization's filing history for revenue/assets/officer pay. Default true. Turn off for a fast name-and-address list.
minRevenue	integer	Drop organizations below this latest reported revenue. 0 = no filter.
minAssets	integer	Drop organizations below this latest reported end-of-year assets. 0 = no filter.
onlyWithFilings	boolean	Skip organizations with no 990 data on record.
maxItems	integer	Max organizations per run (1–1000).

Every search term is combined with every state, so ["food bank"] × ["TX","CA"] runs two queries and merges the results, deduplicated by EIN.

Examples

Qualified prospects — Texas food banks over $1M revenue:

{ "searchTerms": ["food bank"], "states": ["TX"], "minRevenue": 1000000, "maxItems": 100 }

Browse a whole segment — every 501(c)(3) health nonprofit in Wyoming, no keyword needed:

{ "searchTerms": [], "states": ["WY"], "causeArea": "4", "subsectionCode": "3", "maxItems": 200 }

Output (one row per organization)

ein, ein_formatted, name, also_known_as, city, state, address, zipcode, care_of_name, ntee_code, cause_area, subsection_code, organization_type, is_501c3, ruling_date, tax_period, has_filings, filing_count, filing_years, latest_fiscal_year, total_revenue, total_expenses, total_assets, surplus, operating_margin_pct, officer_compensation, revenue_growth_pct, size_tier, irs_revenue_amount, irs_income_amount, irs_asset_amount, latest_filing_pdf, propublica_url, matched_search_term.

Sample row

{
  "ein": "742181456",
  "name": "Houston Food Bank",
  "city": "Houston",
  "state": "TX",
  "ntee_code": null,
  "cause_area": null,
  "organization_type": "501(c)(3) charitable / educational",
  "latest_fiscal_year": 2023,
  "total_revenue": 298504285,
  "total_expenses": 314899919,
  "total_assets": 154916055,
  "surplus": -16395634,
  "operating_margin_pct": -5.49,
  "officer_compensation": 1538220,
  "revenue_growth_pct": -20.69,
  "size_tier": "very large ($50M+)",
  "filing_count": 13
}

Note the null cause_area above: the IRS search index carries no NTEE code for this organization, so the actor reports null rather than guessing a category. Most rows do have one.

Results are sorted largest revenue first when financials are on, so the best-qualified prospects are at the top.

Honest notes

• Financial fields can legitimately be null. An organization can sit in the IRS master file with no Form 990 data at all — usually very small or newly registered ones. Use onlyWithFilings: true to exclude them. Nothing is guessed or filled in.
• latest_fiscal_year is the filing's tax year, not today. IRS 990 data runs a year or two behind by nature; the actor reports the most recent filing that exists.
• Two sets of money fields, on purpose. total_revenue / total_expenses / total_assets come from the organization's own filings. The irs_*_amount fields come from the IRS master file and can differ or lag — both are exposed rather than silently picked.
• minRevenue / minAssets need financials, so setting either switches includeFinancials on automatically and says so in the log.
• Cause area is labelled per row from its own NTEE letter, which is more precise than the numeric group you searched under (group 3 covers both Environment and Animal-Related, for instance).
• Broad queries top out around 1,000 organizations per query — that is the API's practical paging limit, not a cap this actor adds. Narrow by state or cause area to go deeper.
• Requests are paced and retried automatically. An unknown state code is reported as a fixable input problem, never as an outage, and nothing is charged for it.