The Collaborative
Platform for Health AI

Health Universe is the trusted space to publish and run state-of-the-art health AI.

Harvard Medical School UCSF Northeastern University University of Washington Stanford University Carnegie Mellon University

Browse And Launch Health AI Apps With Nothing To Download Or Install

With our cloud platform, you can easily run python apps from your browser, to make research more accessible and actionable. From basic models to state-of-the-art AI. No more time wasted emailing authors, asking for data, or recreating complex software environments.

Our community brings together some of the brightest minds in science and medicine, to accelerate innovation and improve outcomes.

Launch Apps with 1-Click

Health Universe aggregates a wide range of models, from basic science to translational medicine and clinical decision support.

Open Source Collaboration

For too long, health research has been plagued by walled gardens. Break out of silos with open source collaboration.

Advanced ML Support

The community cloud supports basic applications such as clinical calculators all the way to Torch and Tensorflow models.

Clinical Validation

Applications that have been approved for clinical workflows include a validation check and comparative performance metrics.

What this community is doing is brilliant. The impact that you'll have consolidating all these models under one roof is going to transform healthcare. I can't wait to see what happens."

Doug Fridsma

Former Chief Science Officer for the Office of the National Coordinator for Health IT

Finally a group that is taking all these incredible AI models out of acedemia and allowing them to reach physicians at scale, which will fundamentally transform care and outcomes. This is going to be huge."

Carm Huntress
CEO, Credo Health

Frequently asked questions

What is Health Universe?

We are a open-source platform that spans from researchers to clinicians, that is aggregating, organizing, and operationalizing 1-click models, from basic scientific research to state-of-the-art AI. We're a community bringing cutting-edge statistical models to researchers, drug developers, health policy researchers, and clinicians at the point of care to improve outcomes and lower costs. It's the future of health research.

How does it work?

Users contribute by linking a GitHub repository containing a model built in python. Researchers and clinicians can run models directly from Health Universe without having to install any software. The models are made instantly available and interactive using the open source UI library called Streamlit.

How do I know the models are reliable?

Clinical models are meant to inform decision making, but are not meant to be definitive. We support the belief, "humans in charge, AI in the loop." At this stage, machine learning requires human guidance and discretion. Performance metrics are included with each model. Typical metrics are AUC, accuracy, precision and recall. Apps with a "verified" seal have been vetted by pervailing standards.

Who can contribute a model?

Anyone who wants to advance health research. Typical contributors are graduate students, Ph.D students, post-graduates, and industry. Data scientists are also able to contribute by converting existing research into Streamlit compatible projects. We're always looking for new contributors!

What programming languages are supported?

Models must be developed in python and Streamlit. We also support FastAPI for more flexible use cases.

Can't I just use Streamlit Cloud?

Health Universe provides additional features specific to healthcare. We offer researchers and clinicians critical information such as performance metrics, vetting and verification, model origins, SNOMED ontology, EHR integrations, FHIR standards, and more.

How do I contribute my model?

Once you have completed the development of your model, add UI to your project with Streamlit. Then submit your model to Health Universe. We'll spin up a dedicated container to host your model, install your project dependencies, and introduce your model to researchers and clinicians worldwide!

Latest news