CMS Open Payments Scraper

Search CMS Open Payments and export official physician, practitioner, teaching-hospital, manufacturer, amount, date, payment-nature, product, program-year, and source-record data.

The Actor turns bounded searches of the U.S. Centers for Medicare & Medicaid Services (CMS) Open Payments Data API into integration-ready dataset rows.

Use it for Sunshine Act lookup, healthcare compliance review, due diligence, journalism, and repeatable payment-transparency workflows.

What does CMS Open Payments Scraper do?

The Actor queries official General Payment Data published by CMS.

It can search by:

• physician name;
• exact 10-digit NPI;
• manufacturer or GPO name;
• teaching hospital name;
• recipient state;
• payment nature;
• program year; and
• minimum or maximum payment amount.

Every output item represents an individual CMS payment record, not an inferred match or a generated summary.

The Actor discovers current yearly dataset identifiers from the official CMS catalog, so the default latest-year search advances when CMS publishes a new General Payment Data year.

Who is it for?

Compliance teams

Review transfers of value involving a physician, hospital, or reporting entity and feed selected records into a case-management process.

Healthcare due-diligence analysts

Check disclosed industry relationships for a provider or compare records across reporting years.

Researchers and journalists

Export source-linked payment rows for reproducible analysis without manually navigating the Open Payments interface.

Data and automation teams

Schedule searches, persist datasets, connect webhooks, or load the records into a warehouse or spreadsheet.

Why use this CMS Open Payments extractor?

• Official source: requests go to openpaymentsdata.cms.gov.
• No API key: the public CMS Data API is queried anonymously.
• No browser: direct JSON requests reduce runtime and transfer.
• Individual records: results include payer, recipient, amount, date, nature, and record ID.
• Multi-year input: search selected program years in one run.
• Bounded output: maxItems stops collection at the requested limit.
• Typed dataset: common CMS field names are normalized for downstream tools.
• Traceability: every row includes the CMS dataset ID and source endpoint.

What CMS Open Payments data can I extract?

Field	Meaning
recordId	CMS payment record identifier
programYear	Open Payments reporting year
recipientType	Physician or teaching-hospital recipient type
recipientProfileId	CMS recipient profile identifier
recipientNpi	National Provider Identifier when published
recipientName	Combined physician or hospital name
recipientSpecialty	Primary published provider specialty
recipientCity	Recipient business-address city
recipientState	Recipient business-address state
manufacturerName	Manufacturer or GPO making the payment
amountUsd	Total payment amount in US dollars
paymentDate	Date of the payment
numberOfPayments	Transactions represented by the source row
formOfPayment	Cash, in-kind item, dividend, or another CMS form
natureOfPayment	Food, consulting, travel, gift, royalty, or another CMS nature
productName	First associated product, when published
productCategory	First associated therapeutic/product category
contextualInformation	Optional context from the reporting entity
disputeStatus	Whether the record is disputed for publication
sourceDatasetId	Official yearly CMS dataset identifier
sourceUrl	Official CMS Data API endpoint
retrievedAt	UTC retrieval timestamp

CMS fields can be empty when the source does not publish a value for that payment.

How to run a physician payment lookup

1. Open the Actor input page.
2. Enter a physician name or exact NPI.
3. Optionally select program years, state, payment nature, or amount bounds.
4. Set the maximum number of records.
5. Click Start.
6. Open the default dataset to preview or export the payment rows.

A one-word physician name is treated as an exact last name.

A multi-word name uses the first token as the exact first name and the final token as the exact last name.

For the most precise repeatable match, use an NPI.

Input parameters

Input	Type	Required	Description
physicianName	string	conditional	One-word last name or first-and-last physician name
npi	string	conditional	Exact 10-digit NPI
manufacturerName	string	conditional	Case-insensitive manufacturer/GPO name fragment
teachingHospitalName	string	conditional	Case-insensitive hospital name fragment
recipientState	string	no	Two-letter state code
natureOfPayment	string	no	Exact CMS payment-nature label
programYears	string[]	no	Four-digit years from 2019 onward; defaults to latest available
minAmountUsd	number	no	Inclusive minimum amount
maxAmountUsd	number	no	Inclusive maximum amount
maxItems	integer	no	Records to save, from 1 to 10,000; default 100

At least one of physicianName, npi, manufacturerName, or teachingHospitalName is required.

This requirement prevents accidental unbounded scans of datasets containing millions of records.

Input examples

Search by physician name

{
  "physicianName": "Robert Durick",
  "maxItems": 10
}

Search by NPI and state

{
  "npi": "1659344299",
  "recipientState": "OH",
  "maxItems": 25
}

Compare two program years

{
  "npi": "1659344299",
  "programYears": ["2025", "2024"],
  "maxItems": 100
}

Search manufacturer payments above a threshold

{
  "manufacturerName": "Pfizer",
  "minAmountUsd": 100,
  "programYears": ["2025"],
  "maxItems": 500
}

Output example

The default dataset contains normalized JSON such as:

{
  "recordId": "1157125489",
  "programYear": 2025,
  "recipientType": "Covered Recipient Physician",
  "recipientProfileId": "1235813",
  "recipientNpi": "1659344299",
  "recipientName": "ROBERT DURICK",
  "recipientSpecialty": "Dental Providers|Dentist|General Practice",
  "recipientCity": "YOUNGSTOWN",
  "recipientState": "OH",
  "manufacturerName": "ESSENTIAL DENTAL SYSTEMS INCORPORATED",
  "amountUsd": 47.96,
  "paymentDate": "07/24/2025",
  "numberOfPayments": 1,
  "formOfPayment": "Cash or cash equivalent",
  "natureOfPayment": "Gift",
  "productName": "SafeSider HF",
  "productCategory": "DENTAL",
  "disputeStatus": "No",
  "sourceDatasetId": "fb0b1734-1410-429d-92f6-3f4b35218e5e",
  "sourceUrl": "https://openpaymentsdata.cms.gov/api/1/datastore/query/fb0b1734-1410-429d-92f6-3f4b35218e5e/0",
  "retrievedAt": "2026-08-14T20:11:03.056Z"
}

This example reflects current official data. CMS may revise source records in later publications.

How much does it cost to search CMS Open Payments?

Pricing uses one one-time start charge and one item charge per saved payment row.

On the Bronze tier, the start charge is $0.005 per run and each saved item is $0.0096.

Example Bronze prices:

Saved records	Estimated Actor charge
1	$0.0146
10	$0.1010
100	$0.9650
1,000	$9.6050

Higher subscription tiers receive lower per-item prices according to the active Apify pricing table.

No item charge is emitted for duplicate, malformed, or rejected source rows.

Apify platform usage may be shown separately according to your plan.

Scheduling compliance monitoring

Create an Apify Schedule with a stable NPI or company search.

For year-over-year review:

1. keep a versioned copy of each run dataset;
2. use a consistent NPI or exact physician identity;
3. include the relevant program years;
4. compare on recordId and source fields; and
5. treat CMS corrections as source revisions, not automatically as new payments.

The Actor does not maintain a private historical database or send alerts itself.

Use Apify webhooks, integrations, or your own workflow to react to completed runs.

Export and integration workflows

You can export the default dataset as JSON, JSONL, CSV, XML, RSS, or Excel through Apify dataset endpoints.

Common workflows include:

• append NPI-specific rows to a compliance warehouse;
• send high-value payment rows to a review queue;
• compare selected program-year exports in a notebook;
• load manufacturer results into a spreadsheet;
• trigger a webhook after a scheduled run; or
• retrieve output from Make, Zapier, n8n, or another integration.

Use recordId as the source identifier and retain programYear plus sourceDatasetId for audit context.

Run with the Apify API using cURL

Replace YOUR_TOKEN with your Apify API token.

curl -X POST \
  "https://api.apify.com/v2/acts/automation-lab~cms-open-payments-data/runs?token=YOUR_TOKEN&waitForFinish=120" \
  -H "Content-Type: application/json" \
  -d '{"npi":"1659344299","maxItems":10}'

Fetch dataset items using the defaultDatasetId returned by the run API.

Run with JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });

const run = await client.actor('automation-lab/cms-open-payments-data').call({
  npi: '1659344299',
  programYears: ['2025', '2024'],
  maxItems: 100,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

Run with Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_TOKEN")
run = client.actor("automation-lab/cms-open-payments-data").call(
    run_input={"physicianName": "Robert Durick", "maxItems": 10}
)
items = client.dataset(run["defaultDatasetId"]).list_items().items
print(items)

Use with MCP and AI agents

Add the Apify MCP server to Claude Code:

claude mcp add --transport http apify \
  "https://mcp.apify.com?tools=automation-lab/cms-open-payments-data"

Claude Desktop setup

Cursor setup

VS Code setup

Each client can use the same remote MCP server configuration (place it in the configuration location required by that client):

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com?tools=automation-lab/cms-open-payments-data"
    }
  }
}

Example prompts:

• “Search CMS Open Payments for NPI 1659344299 and return the ten newest available payment rows.”
• “Export this physician's 2024 and 2025 CMS payment records for a compliance review.”
• “Find CMS General Payment Data for this manufacturer above $500 and summarize it in a table.”

Verify sensitive compliance conclusions against the linked official CMS record and your organization's policies.

Limits and source behavior

CMS publishes General Payment Data by program year and can correct records in later publication cycles.

The Actor supports program years present in the live CMS catalog, currently from 2019 onward.

Results follow the order returned by the official API; they are not claimed to be sorted by amount or date.

A name search is exact on parsed first/last tokens, while manufacturer and hospital searches use name fragments.

Very broad manufacturer or hospital fragments may take longer than precise NPI or physician queries.

The Actor retries transient network, HTTP 429, and server failures with bounded backoff.

Malformed input, unavailable program years, stable client errors, and invalid CMS response shapes fail the run rather than silently returning misleading empty output.

A valid query with no matching source records succeeds with an empty dataset.

Data quality and interpretation

Open Payments reports transfers of value submitted under federal program rules.

A listed payment does not by itself establish misconduct, bias, or a financial conflict.

Names can collide, addresses can change, and some fields can be absent.

Prefer NPI or CMS profile identifiers for identity-sensitive work.

Preserve source IDs and timestamps when using records as compliance evidence.

Consult CMS documentation and qualified legal or compliance professionals for interpretation requirements.

Responsible use

The Actor accesses public U.S. government data without bypassing access controls.

You are responsible for lawful use, retention, sharing, and interpretation of the output.

Do not use payment data as the sole basis for an adverse employment, credentialing, insurance, healthcare, or legal decision.

Follow applicable privacy, anti-discrimination, due-process, and records-management obligations.

Troubleshooting

Why did my run return no records?

Check spelling, NPI, selected year, state, amount bounds, and exact payment-nature label.

Try an NPI-only search to reduce name ambiguity.

Remember that omission of programYears searches only the latest available General Payment Data year.

Why was my input rejected?

At least one primary search field is required.

NPI must have exactly ten digits, state must have two letters, years must have four digits, and amount bounds cannot be negative or reversed.

Why did a CMS request retry?

The public CMS API can respond slowly or transiently fail.

The Actor retries a bounded number of times and logs concise reasons; persistent failures correctly fail the run.

Why do two years have different fields or values?

CMS publication content can evolve and reporting entities can submit corrections.

The Actor normalizes a stable subset and leaves unavailable values as null.

FAQ

Does the Actor need a CMS API key?

No. It uses public official endpoints.

Does it use a proxy or browser?

No. It makes direct JSON requests and has no automatic paid-proxy fallback.

Does it include research and ownership payments?

No. This release searches individual General Payment Data records. Research and ownership datasets are outside the current contract.

Can it search by NPI?

Yes. NPI is the most precise supported provider input.

Can I search several years?

Yes. Supply programYears as four-digit strings, for example ["2025", "2024"].

Can it monitor changes automatically?

Schedule repeated Actor runs and compare stored datasets or use a webhook. The Actor extracts records but does not itself retain history or issue alerts.

Is this medical or legal advice?

No. It supplies public records for analysis and does not provide medical, legal, or compliance conclusions.

Related automation-lab Actors

For adjacent public-data workflows, explore the automation-lab Actor portfolio.

Only use related Actors whose source and output contract match your actual workflow; CMS Open Payments records are not a substitute for licensing, sanctions, clinical, or credential data.

Support

If a run fails unexpectedly, include the run URL, a redacted input, the affected program year, and whether the same query works on the official CMS Open Payments site.

Do not include Apify tokens, private case notes, or unrelated personal information in a support report.