CMS Provider Facilities Scraper (Medicare)

What this Actor scrapes

The CMS Provider Data Catalog (data.cms.gov/provider-data) is the federal government's authoritative directory of every Medicare-certified healthcare facility in the country — nursing homes, hospitals, home health agencies, and hospices — complete with address, phone, ownership, staffing, and the CMS 1-5 star quality ratings that drive Medicare's own "Care Compare" tool.

The catalog's own datastore-query API is public, but turning it into a reliable bulk export takes real engineering: each facility type publishes a *different* column set (a hospital record and a nursing-home record share almost no field names), the national datasets run past 450 000 rows, and a naive client either downloads entire CSVs or has to reverse-engineer the datastore's bracketed conditions[] filter syntax to query server-side. We did that reverse-engineering, curated four facility datasets behind one friendly dataset input, and hand you back one richly-typed row per facility — core columns normalised, everything else preserved in a typed extras map.

What we handle for you

• 🔁 Real limit/offset pagination — the datastore API caps nothing on paper, but we page in bounded 500-row chunks and stop exactly when the server-reported match count is reached, so a broad query doesn't run away.
• 🛡️ Retry with exponential backoff — on 408 / 429 / 5xx we back off (2s → 30s, capped) and honour Retry-After, up to 5 attempts per page, before surfacing an error.
• 🧊 Per-item fault isolation — one malformed record in a page of 500 is logged and skipped; it never sinks the rest of the page.
• 🧱 Pydantic-validated rows — a stable core (CCN, name, address, city, state, ZIP, phone) plus a typed extras map for every dataset-specific column (star ratings, ownership, bed counts, services offered). Nothing degrades to a bare, untyped dict.
• 🌐 Proxy-ready — data.cms.gov is a public federal endpoint and direct routing works on most infrastructure; the Apify Proxy field is there when your network needs it.
• 💰 You pay only for results that land. No data → no charge beyond the small run warm-up fee.

Use cases

• Healthcare vendor sales lists — pull every nursing home or hospital in a state for a targeted outreach or prospecting list.
• ESG / site-diligence screening — check quality ratings, ownership type, and inspection history before an acquisition or partnership.
• Compliance & referral-network products — bulk-export by facility type and resell structured datasets to billing, credentialing, or referral tools.
• Journalism & policy research — analyze facility quality, ownership concentration, or staffing trends by state.
• Competitive intelligence — track how a chain's facilities rank across the CMS star-rating system.
• Home health / hospice referral sourcing — build a state-scoped directory of agencies with services offered and quality scores.

Input

Paste this into the Apify Console, or send it as the run input over the API. Proxy settings are on by default; you rarely need to touch them.

{
  "dataset": "nursing_homes",
  "state": "CA",
  "maxResults": 100,
  "proxyConfiguration": {
    "useApifyProxy": false
  }
}

Output

One row per result, schema-validated before it is written. Export JSON, CSV, Excel or XML from the run, or read it over the API.

datasetccnnameaddresscitystatezip_codephoneextrasscraped_at

{
  "dataset": "nursing_homes",
  "ccn": "015009",
  "name": "BURNS NURSING HOME, INC.",
  "address": "701 MONROE STREET NW",
  "city": "RUSSELLVILLE",
  "state": "AL",
  "zip_code": "35653",
  "phone": "2563324110",
  "extras": {
    "overall_rating": "2",
    "ownership_type": "For profit - Corporation",
    "number_of_certified_beds": "57"
  },
  "scraped_at": "2026-08-26T00:00:00+00:00"
}

Pricing

You pay only for results that land. Cap any run with maxTotalChargeUsd. See pricing & billing for worked examples.

Limitations

• Column sets differ per dataset — a hospital record and a nursing-home record share almost no raw field names. The stable core (CCN, name, address, city, state, ZIP, phone) is normalised; everything else lands in extras under its original CMS field name, which varies by dataset.
• No email addresses — the Provider Data Catalog does not publish facility email addresses. Phone is included; email is not part of the public catalog.
• US facilities only — the catalog covers Medicare-certified US facilities. International facilities are out of scope.
• Data freshness — CMS refreshes each dataset on its own cadence (often quarterly for quality ratings). The scraped_at field marks when the Actor pulled it, not when CMS last updated the underlying record.
• 5000-result hard cap per run — for a full-state pull larger than that, split by city or run multiple jobs and merge the datasets.