Data licence and attribution
Last updated 2026-08-23.
AINPI publishes three different things, and they are not under the same terms. This page says which is which, because a data consumer needs to know before they build on it.
The federal source data: public domain
The underlying files come from the US government: the CMS National Provider Directory bulk export, NPPES, PECOS enrollment files, the OIG List of Excluded Individuals and Entities, the SAM.gov exclusions extract, CMS Provider Data Catalog files, USDA ERS county codes and US Census geography.
Works of the United States government are not subject to copyright protection in the United States. AINPI claims no rights over that source data and grants none, because it has none to grant. You do not need permission from AINPI to use it, and nothing on this page restricts what you may do with the federal files themselves.
Each source carries its own terms of use at the publisher. Where a source is not a US government work, the data sources page names it and links its terms.
The AINPI code: Apache-2.0
Every ingestion, analysis and site script is released under the Apache License 2.0. The full text is in the repository. Fork it, run it, change it, sell what you build with it.
The AINPI compilation and findings: Apache-2.0, attribution requested
The derived layer is the part that took work: the flattened tables, the
release archive, the cross-source crosswalks, the per-state audit slices and
the finding payloads under /api/v1/. That work is joining federal files to
each other and to vendor files, and reconciling the result.
It is published under the same Apache-2.0 terms as the code, which permits commercial use, redistribution and modification.
Attribution is requested rather than required by that licence, and we ask for it anyway:
Source: AINPI (ainpi.dev), derived from CMS National Provider Directory public use files.
If you redistribute a modified version, please say that you changed it, so a reader can tell your numbers from ours.
What attribution is for
It is not credit for its own sake. Every number here is tied to a release date and a methodology version, both of which change. A figure quoted with no provenance cannot be checked against the release it came from, and this project has already had to correct itself when an upstream source changed shape underneath a published claim. Naming the source and the release is what makes the correction reachable.
Cite the release, not just the site:
AINPI, "The role gap is a Medicare-billing gap", NDH release 2026-08-20, methodology 0.7.2-draft. https://ainpi.dev/findings/role-gap-composition
The release archive
CMS serves only the current version of its bulk export. AINPI keeps the earlier
ones and republishes them unchanged apart from format conversion, plus a
release_date column identifying which release each row came from.
The archive is federal public-domain data. It is free, and it is intended to stay free.
One thing to know before you diff two releases. Practitioner and
Organization ids embed an NPI and are stable across releases. Endpoint and
Location ids are random UUIDs that CMS regenerates on every export, so joining
those two on _id across releases reports 100% churn that did not happen. Join
endpoint on its address instead. The table comments carry the same warning.
No warranty
Nothing here is a warranty of accuracy, currency or completeness. See the terms of use, particularly the section on verifying against primary sources before acting on any record about a named provider.