Explore datasets for healthcare research.
Earn from your data while patient records stay in-house.
Build and benchmark better healthcare AI.
Supported by
01 · Our mission
That work showed us what's broken in healthcare research: the data exists, and almost nobody can use it. Researchers wait months for access, and by the time the data arrives, the grant deadline has passed or the data isn't what they needed.
Privacy infrastructure protects patients for very real reasons, but it also locks data away from the people who could use it to advance medicine. As frontier AI improves, access to high-quality clinical data will matter as much as the models themselves.
So we generate digital twins of healthcare datasets: synthetic replicas that preserve the statistical structure of real patient records without carrying over any individual patient's record. Health systems hold an enormous amount of medical knowledge, and most of it goes unused.
of healthcare data is never used for research
wait for access to health data for a complex research request
of FDA-authorized AI medical devices have no published clinical validation data
02 · Who it's for
GenMD turns clinical data locked inside health systems into privacy-preserved digital twins, so researchers, data providers, and AI teams can each put it to work.
Explore privacy-preserved digital twins of real clinical datasets.
Start analysis in days, not the 6 to 12 months data access usually takes.
Validate findings on real data, with institutional sign-off.
Explore the datasets
Create a new revenue stream from data your institution already holds.
GenMD deploys in your own cloud account. Real patient data stays with you.
The right to decline any partner, a full audit trail, and transparent terms.
Put your data to work
Train and evaluate models on high-quality, real-world clinical data twins.
Benchmark frontier models on clinical cases, with health system partners.
De-identified twins, tested for re-identification risk.
Explore AI partnerships
03 · How it works
Each step runs under the health system's governance. Real patient data stays inside it.
01
Partner
A health system joins and sets its terms, including the right to decline any partner.
02
Connect
GenMD deploys in the institution's own cloud account.
03
Twin
We generate a digital twin of the dataset and validate its privacy and fidelity.
04
Deliver
Vetted researchers and AI teams work on the twin. Results can be validated against real data inside the health system.
05
Earn
When a partner licenses access to the real data, the institution keeps the majority of the revenue.
04 · The data
One of the many datasets we've released on Healthcare Datahub: de-identified medical and pharmacy claims, linked at the patient level and ready for research.
Featured dataset
Medical & Pharmacy Claims
Claim records
Distinct patients (est.)
Distinct encounters
Activity window
States, by provider location
Documented fields
Coded with
ICD-10
ICD-10-PCS
CPT
HCPCS
NDC
NPI
MS-DRG
APC
For researchers & AI teams
Clinical data you can actually use.
Privacy-preserved datasets, available in days instead of months.
Explore the data
For health systems
Revenue from data you already hold.
See every query, decline any partner, and share in the revenue.
Become a partner
Privacy-preserving clinical data for research and frontier AI.