GuideStar / ProPublica Nonprofit Scraper

Find public US nonprofit profiles by EIN or search phrase and return structured organization data, IRS Form 990 summary metrics, filing history, and source links. The actor is designed for grant researchers, nonprofit due-diligence teams, journalists, fundraisers, and AI agents that need a small, traceable nonprofit dataset without a login.

This actor uses the public ProPublica Nonprofit Explorer API v2. ProPublica organization objects can include a guidestar_url, so the output preserves that GuideStar profile link when available. GuideStar is now part of Candid; this actor does not authenticate to Candid, bypass access controls, or claim to return paid GuideStar/Candid-only fields.

What it extracts

One dataset item represents one nonprofit identified by EIN.

Field group	Included data
Identity	EIN, formatted EIN, organization name, alternate name, address, city, state, ZIP
Tax classification	IRS subsection code, NTEE code, organization status, classification, deductibility, ruling date
Financial snapshot	Latest filing year/type, revenue, expenses, assets, liabilities
Filing history	Up to maxFilingsPerOrganization compact Form 990 / 990-EZ / 990-PF summaries
Public profiles	ProPublica, GuideStar, and NCCS URLs when supplied by the source
Provenance	Search query, API search URL, source identifier, scrape timestamp, warnings

The API source is a public index of IRS-exempt-organization and Form 990 data. A missing value is returned as null, not invented or silently removed. Filing fields vary by form type, so the actor normalizes common financial aliases and keeps the result shape stable.

When to use it

Use exact EINs when you already have a list of organizations and need their public profiles. Use searchQueries when you want discovery by organization name, alternate name, or city. Add state, nteeCode, or subsectionCode to narrow a search. Keep includeFilings enabled when you need financial metrics; disable it for faster profile-only discovery.

Do not use this actor for private donor data, Candid account data, paid GuideStar reports, real-time charity ratings, grant-award databases, or legal/tax determinations. It is also not a general-purpose web search tool.

Example inputs

Small discovery run:

{
  "searchQueries": ["food bank"],
  "state": "NY",
  "subsectionCode": 3,
  "maxResults": 25,
  "includeFilings": true,
  "responseFormat": "detailed"
}

Exact EIN lookup:

{
  "eins": ["14-2007220", "13-1837418"],
  "maxResults": 2,
  "maxFilingsPerOrganization": 3
}

For an AI-agent response with fewer nested fields:

{
  "searchQueries": ["community health"],
  "state": "CA",
  "maxResults": 10,
  "responseFormat": "concise"
}

Output example

{
  "ein": "142007220",
  "formattedEin": "14-2007220",
  "name": "PRO PUBLICA",
  "city": "NEW YORK",
  "state": "NY",
  "subsectionCode": 3,
  "nteeCode": "A20",
  "latestFilingYear": 2024,
  "latestFilingType": "Form 990",
  "latestRevenue": 1035696,
  "latestExpenses": 983566,
  "filingCount": 8,
  "filings": [],
  "guidestarUrl": "https://www.guidestar.org/organizations/14-2007220/.aspx",
  "proPublicaUrl": "https://projects.propublica.org/nonprofits/organizations/142007220",
  "source": "propublica-nonprofit-explorer",
  "warnings": []
}

When responseFormat is detailed, filings contains the newest selected filing summaries. In concise mode it is an empty array while latest filing metrics and filingCount remain available. See the dataset schema for the exact stable shape.

Pricing

Event	Price
Actor start	$0.00005, scaled by allocated memory
Stored nonprofit record	$0.005 each

The actor charges a record only after its dataset item is stored. A run with maxResults: 25 has a maximum record-event estimate of $0.125, plus the actor-start event and normal Apify platform usage. The run status and OUTPUT key-value record report collected records, charged events, request errors, and warnings.

Reliability and source limits

The actor uses the documented ProPublica API v2 search endpoint, which returns 25 organizations per page, and the EIN organization endpoint for filing detail. It retries transient request failures with a short backoff, paces requests with requestDelayMs, deduplicates EINs, persists progress, and returns a profile-only record with a warning if filing detail fails. Empty valid searches are successful and clearly reported.

ProPublica notes that the API is subject to change. If fields disappear or a request begins failing, provide the run ID, source URL, and an example EIN in an actor issue. The API and IRS data are not a substitute for verifying a nonprofit with the source agency or a qualified professional.

JavaScript and Python API usage

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('YOUR_USERNAME/guidestar-propublica-nonprofit-scraper').call({
  eins: ['14-2007220'],
  includeFilings: true,
  responseFormat: 'detailed'
});
console.log(run.defaultDatasetId);

from apify_client import ApifyClient

client = ApifyClient(token='YOUR_APIFY_TOKEN')
run = client.actor('YOUR_USERNAME/guidestar-propublica-nonprofit-scraper').call(
    run_input={'searchQueries': ['food bank'], 'state': 'TX', 'maxResults': 10}
)
print(run['defaultDatasetId'])

The input and output are also suitable for Apify's hosted MCP integration: the input descriptions route an agent to exact EIN lookup versus discovery, and each dataset item has semantic field titles, descriptions, and provenance.

Responsible use

This independent actor is not affiliated with ProPublica, GuideStar, Candid, the IRS, or NCCS. Respect the source's data terms, robots and rate limits, applicable privacy and data-protection law, and any restrictions on downstream outreach. Treat public financial figures and tax classifications as research data; verify material grant, compliance, eligibility, and tax decisions against current primary records.

Official source documentation

• ProPublica Nonprofit Explorer API
• Candid developer portal
• Candid notice about decommissioned GuideStar Gen 1 APIs