US Healthcare Provider Search: Normalized NPI Data

Get every healthcare provider matching your criteria, from the complete US NPPES registry (9.4M+ providers), as clean rows you can use immediately.

True statewide pulls. No 1,200-result cap.

Most NPI tools on Apify and elsewhere are thin wrappers around the live NPPES API, which returns at most 200 records per request and 1,200 results per query, hard cap. "Every cardiologist in California" is roughly 5,000 providers. A wrapper cannot deliver that list. This actor is not a wrapper: it queries our own database, built from CMS's full bulk data files, so a 50,000-row statewide pull is just a bigger run.

Several competing listings say this themselves: their advice for a statewide pull is to split your search into ZIP-by-ZIP runs and merge the datasets yourself. That workaround is your time and your money. Here it is just a query.

Raw NPPES is messy. You are paying us to have cleaned it.

The government registry has no input sanitization. Real examples of what that means, and what this actor returns instead:

Raw NPPES	This actor
credential = "MSN, APRN, FNP-C" (free text, 57% of records blank)	primary_credential = "FNP" plus every credential parsed and ranked. Filter MDs without also getting the MPH holders.
"123 MAIN ST STE 200" and "123 Main St, Suite 200" as two different addresses	One canonical, deduplicated address with parsed components and a PO Box flag
Endpoints, secondary locations, and alternate names shipped as three separate files you must join yourself	One provider record with everything pre-joined
Taxonomy code 207RH0003X	"Internal Medicine, Hematology & Oncology", plus our own 21-bucket buyer-oriented grouping

CMS's own audit programs have found 45-49% of provider directory listings contain inaccuracies, and only about 8% of NPI records are updated by their providers in a given year. No vendor can fix what providers never report. What we do about it: refresh from CMS weekly, deduplicate and normalize every field, attach a 0-100 quality confidence score to every record so you can filter to high-confidence rows, and stamp every response with the exact NPPES publication date it came from. You always know how fresh your data is.

Included on every record (no add-on fees)

• OIG exclusion flag: is this provider on the HHS exclusion list (LEIE)? Avoid mailing sanctioned providers.
• Medicare enrollment (PECOS) flag: set the medicare_enrolled input filter to pull "Medicare-accepting providers only."
• Practice-group clustering: pull the whole cardiology group at one address in one query.
• Multi-state licensure: providers licensed in 3+ states, a telehealth-ready signal.
• Digital maturity flags: has a FHIR endpoint, has Direct messaging.
• Career stage and newly-enumerated flag: find brand-new practices.
• Quality confidence score (0-100) with per-component breakdown.

Other services sell some of these flags separately. What you cannot get elsewhere at usage pricing is the bundle: the flags integrated into normalized, deduplicated records with parsed credentials, priced per result instead of per month.

Filters

State, specialty (taxonomy classification), credential code, city, ZIP, entity type (individual/organization), and name prefix (first, last, or business name). Set max_results to control spend. Set include_full_record for the complete enriched record per provider (nested taxonomies, all credentials, locations); leave it off for a compact row.

Use from AI agents (MCP)

This actor works as an MCP tool out of the box. Point your agent at mcp.apify.com with this actor enabled and it can run provider searches directly ("find every dermatology NP in Texas"). Results land in a dataset your agent can page through.

Pricing

$3 per 1,000 results ($0.003 per row pushed to your dataset). No subscription, no minimum. You pay only for rows delivered: duplicates and errors are filtered before they reach your dataset. A 10,000-provider statewide specialty pull costs $30.

Stability

CMS changes its file formats (a major version cutover happened in March 2026 and broke downstream loaders across the ecosystem). We absorb that churn behind a stable output schema: your integration does not change when CMS's files do.

Data source and compliance

Data comes from CMS's public NPPES registry (FOIA-disclosed federal data on healthcare providers, published for exactly this kind of use) plus other public federal datasets (HHS OIG LEIE, CMS PECOS). This is provider enrollment data. It contains no patient data and no HIPAA-covered information. Records are served as published by CMS, normalized for usability; addresses and phones are as reported by providers to CMS, not independently verified.