Data · Public Sources
35 documented pages. A selected public-source catalog.
35 dataset pages with source and freshness metadata where the catalog supplies it. FHIR R4 covers the supported provider resources; provenance fields can be null.
| Dataset | Agency | Catalog observation |
|---|---|---|
| Daily Nurse Staffing | CMS | 2025-06-30 |
| OIG Exclusions | HHS-OIG | 2026-05-08 |
| SNF Ownership Records | CMS | 2026-04-15 |
| Care Compare Nursing Homes | CMS | 2026-05-07 |
| Care Compare Home Health | CMS | 2026-05-07 |
| Care Compare Hospice | CMS | 2026-05-07 |
| Care Compare Dialysis | CMS | 2026-05-07 |
| Provider of Services | CMS | 2026-05-07 |
Published cadence
Catalog entries state an intended source cadence; dated observations show what production actually holds.
Available provenance
The schema can carry source, snapshot date, methodology, and confidence. Population varies by dataset and export.
FHIR R4 resource
FHIR R4 covers the supported Practitioner, Organization, Location, PractitionerRole, and HealthcareService resources.
How it works
Federal data → field-level provenance → your agent.
Catalog responses identify the source and observation metadata available for a row. Some fields lack a complete provenance tuple; missing values stay null rather than being filled synthetically.
01
Federal source pull
Registered jobs pull from CMS, HHS-OIG, HRSA, BLS, and BEA. The source's declared cadence does not prove the latest successful load; check the displayed loaded date. Available attestation metadata applies to specific snapshots, not every pull or row.
02
Source and capture context
The nullable schema can carry source, snapshot date, methodology version, confidence tier, and other fields. Inspect each export or API response for the elements actually populated. Bulk availability is listed separately in the machine-readable manifest.
03
FHIR R4 + manifest-led CSV
Supported provider resources are available through FHIR R4. Bulk files appear only in the manifest after their delivery format and schema are checked; a checksum applies only when the listed response actually supplies one.
All datasets
35 dataset pages.
CMS, HHS-OIG, HRSA, and NIH sources. Nursing facilities, provider performance, sanctions, hospital financials, care quality, workforce and access, and research activity.
| Daily Nurse Staffing | CMS | Nursing facilities | 2025-06-30 |
| OIG Exclusions | HHS-OIG | Sanctions and exclusions | 2026-05-08 |
| SNF Ownership Records | CMS | Ownership networks | 2026-04-15 |
| Care Compare Nursing Homes | CMS | Nursing facilities | 2026-05-07 |
| Care Compare Home Health | CMS | Care quality | 2026-05-07 |
| Care Compare Hospice | CMS | Care quality | 2026-05-07 |
| Care Compare Dialysis | CMS | Care quality | 2026-05-07 |
| Provider of Services | CMS | Nursing facilities | 2026-05-07 |
| Hospital Cost Reports | CMS | Hospital financials | 2026-05-01 |
| Open Payments | CMS | Provider performance | 2026-04-30 |
| QPP Individual Scores | CMS | Provider performance | 2026-05-24 |
| Care Compare ASCs | CMS | Care quality | 2026-05-07 |
| PECOS Ordering & Referring | CMS | Provider performance | 2026-06-14 |
| Hospital Price Transparency Enforcement | CMS | Hospital compliance | 2026-06-06 |
| HRSA Health Professional Shortage Areas | HRSA | Workforce and access | 2026-06-14 |
| HRSA Medically Underserved Areas / Populations | HRSA | Workforce and access | 2026-06-14 |
| HRSA 340B Covered Entities | HRSA | Workforce and access | 2026-06-14 |
| Medicare Shared Savings Program ACO Participants | CMS | Provider performance | 2026-06-14 |
| ClinicalTrials.gov Studies | NLM | Research activity | 2026-06-14 |
| NIH RePORTER Projects | NIH | Research activity | 2026-06-14 |
| FDA Enforcement Reports | FDA | Drug and device safety | 2026-06-14 |
| Drugs@FDA Approved Products | FDA | Prescription drugs | 2026-06-14 |
| FDA National Drug Code Directory | FDA | Prescription drugs | 2026-06-14 |
| FDA Device Establishment Registration | FDA | Drug and device safety | 2026-06-14 |
| Medicare Opt Out Affidavits | CMS | Provider performance | 2026-06-14 |
| Medicare Revalidation Due Dates | CMS | Provider performance | 2026-06-14 |
| Medicare Physician & Other Practitioners by Service | CMS | Provider performance | 2026-06-14 |
| CMS Market Saturation & Utilization | CMS | Workforce and access | 2026-06-14 |
| Medicare Part D Spending by Drug | CMS | Prescription drugs | 2026-06-14 |
| NUCC Health Care Provider Taxonomy | NUCC | Reference data | 2026-06-14 |
| SAM.gov Exclusions | GSA | Sanctions and exclusions | 2026-06-13 |
| CMS Civil Money Penalties | CMS | Sanctions and exclusions | 2026-07-12 |
| OIG Corporate Integrity Agreements | HHS-OIG | Sanctions and exclusions | 2026-06-13 |
| HRSA Uniform Data System (UDS) | HRSA | Workforce and access | 2026-06-15 |
| State Professional Licensing & Disciplinary Records | STATE | Workforce and access | 2026-06-15 |
Nullable provenance schema
Up to 14 provenance fields.
A supported record can expose up to fourteen provenance fields. Population varies by source, field, endpoint, and export; absent source, date, or methodology values remain null.
Federal source pull
A source snapshot can record a date and SHA-256 for the identified object. Inspect the specific record before treating a digest as a raw-file hash.
Column → field mapping
Registered source columns use source-specific typed mappings. A mapping version appears in _methodology where populated.
Source and capture context
Supported fields can expose _source, _snapshot, _confidence, and other nullable provenance elements. FHIR and CSV surfaces document their own metadata coverage; the bulk manifest separately identifies the files available for download.
Why it's defensible
The moat compounds with every source family.
Federal data is public. The moat is the engineering layer on top of it — provenance depth, schema consistency, and FHIR compatibility that only compound over time.
Cross-source joins
CCN identity backbone (CMS POS) lets us join nursing home, dialysis, hospice, and staffing data deterministically. Competitors who start fresh pay 18+ months to get there.
Methodology versioning
Every column mapping is a versioned artifact. When CMS changes a schema, we ship a new methodology version — prior snapshots remain traceable to the version that produced them.
FHIR R4 from day one
USCDI v3 resource mapping for all five provider resource types. Agents and EHRs that speak FHIR get a zero-translation integration path.
Public-domain foundation
U.S. Government Works upstream. No NMLS. No state bar. No ABMS. No data that can be yanked by a licensor. The raw data is yours too — what you're licensing is the provenance layer.
Get access
A selected public-source catalog within the active production registry (registry status is not completeness, load status, or freshness). Nullable provenance. Ready for your agent.
Pilot access is operator-approved. Reviewed within one business day. No charge during the pilot period.
What the Fonteum graph holds, by domain.
Sources integrated, sources with data, and fresh source observations are separate lifecycle counts from the dated database snapshot. No platform-wide source-completeness count is published. Completeness is source-specific and must be evaluated against the named source's expected scope. State Medicaid jurisdiction coverage is a separate measure. The observation does not mean every upstream dataset refreshed that day. Metrics marked ≈ use O(1) planner estimates; unmarked metrics use exact queries. Re-check the full set any time at /api/v1/platform-stats.
Provider graph
Ingested entity records and the source-linked field facts behind them.
- Sources integrated111
- Fonteum source registry
- Sources with data90
- Fonteum snapshot ledger
- Fresh source observations73
- Fonteum snapshot ledger
- Source-linked field facts≈26.2M
- Fonteum provenance ledger · 26,211,194 rows
- Ingested entity records≈13.4M
- Fonteum entity graph · 13,393,315 rows
- NPPES NPI records≈9M
- CMS NPPES (NPI registry) · 9,016,856 rows
Payments & prescribing
CMS Medicare Part D prescribing and Open Payments industry-transfer records.
- Part D prescriber records≈27.2M
- CMS Medicare Part D Prescribers · 27,230,788 rows
- Open Payments transactions≈16.2M
- CMS Open Payments · 16,168,159 rows
Enforcement & exclusions
Federal and state exclusion, debarment, and sanction records.
- State Medicaid jurisdictions13
- Fonteum state-exclusions serving table
- Federal exclusion records324,126
- SAM.gov exclusions (GSA)
- Excluding agencies56
- SAM.gov exclusions (GSA)
- OIG-LEIE exclusions83,464
- HHS-OIG LEIE
- State Medicaid exclusions22,917
- State Medicaid exclusion lists
Licensing
State professional-license rosters.
- State professional licenses≈8.1M
- State licensing boards (multi-state) · 8,133,816 rows
Research corpus
Clinical-trial records and NIH-funded projects.
- Clinical trials589,453
- ClinicalTrials.gov
- NIH research projects179,957
- NIH RePORTER
/sources → the generated source-catalog subset, with documented agency, publisher cadence, loaded-date context where available, and license.
Global registers
UK Companies House and Australia ASIC register records are available at /global, with a partial Corporations Canada subset, page-level source dates and license attribution; field-level source coverage varies. The ABN definition explains why an Australian Business Number is not the same as an ASIC company number. Available archive formats and their limits are documented in the bulk-export guide. The ASIC serving table contained 3,952,491 rows when queried 2026-08-12.