Medicare US Scraper

Medicare US Scraper extracts structured US healthcare provider data from the official CMS Medicare Physician & Other Practitioners dataset. Use it for provider lookup, geographic filtering, provider-type analysis, NPI-driven workflows, and lightweight health checks.

This actor is not affiliated with CMS or Medicare. It uses public data from data.cms.gov and does not require proxies.

Supported operations

• probe: checks CMS Medicare dataset availability with a stats request.
• categories: lists browse categories for states and provider types.
• filters: returns filter metadata and sort options.
• search: searches the CMS Medicare dataset by keyword or exact filters.
• category: browses providers by category_id in state:CA or provider_type:Cardiology format.
• item: returns a detailed provider profile with beneficiary and clinical data.

Input guide

operation is required. For search, provide query or at least one exact filter such as npi, first_name, last_name_or_org, city, state, zip_code, provider_type, entity_type, or medicare_participating.

For item, provide npi or item_id (alias for NPI).

limit defaults to 25 and is capped at 100. page is 1-based. include_total defaults to true and controls whether an extra stats request is made for exact counts.

Example inputs

Probe:

{
  "operation": "probe"
}

Search:

{
  "operation": "search",
  "query": "cardiology",
  "limit": 5
}

Search with exact filter:

{
  "operation": "search",
  "state": "CA",
  "provider_type": "Cardiology",
  "limit": 5
}

Category browse:

{
  "operation": "category",
  "category_id": "state:NY",
  "limit": 5
}

Item detail:

{
  "operation": "item",
  "npi": "1003000126"
}

Output

For monetization and result counting, each useful result is written as its own Apify dataset item. Search limit 5 writes 5 dataset items, and item lookup writes 1 dataset item when the provider is found.

Run status, totals, no-result summaries, and error summaries are written to the default key-value store under OUTPUT. Probe, no-result searches, invalid item lookups, and invalid inputs do not create fake dataset result rows.

Search and category results can include:

• status, source, operation, mode, record_type
• id, item_id, npi, slug
• name, first_name, last_name_or_organization, credentials
• entity_type, provider_type, medicare_participating
• address, city, state, zip, country
• ruca_code, ruca_description
• totals (beneficiaries, services, charges, amounts)
• activity_summary

Item results extend listing fields with:

• dataset_id, dataset_name, landing_url
• address_lines
• beneficiary_profile (age, risk, demographics)
• service_mix (drug and medical services)
• clinical_profile (condition percentages)
• source_record

No-result searches write zero dataset items and return status: "success" with dataset_item_count: 0 in OUTPUT. Invalid item lookups write zero dataset items and return status: "not_found" with a structured error object in OUTPUT.

Limitations

• Data freshness, coverage, and field availability are controlled by CMS.
• This is a Medicare provider summary and utilization dataset, not a live appointment directory.
• The CMS public API may rate-limit heavy usage.
• Stats endpoint for total counts may be temporarily unavailable (actor falls back gracefully).
• This actor does not provide medical advice and should not be used as a substitute for official provider verification.

Proxy behavior

No proxy is required. The actor performs direct HTTPS requests to public CMS endpoints.