Note-to-FHIR

New release supporting Dutch - available now!

Read more
 
 
 

The only Open Platform for Generative AI in Healthcare

Generative AI will reshape healthcare. It holds the enormous potential to revolutionize the everyday care journeys of healthcare professionals and their patients. To trust such a new and transformative technology, it needs to be safe, secure and responsible. That is why we have developed an open source generative AI platform fully dedicated to healthcare. Hosting our own Large Language Model as well as other models, applications and plug-ins. Always on top of the latest developments, always state of the art. And Always Open.

We encourage collaboration between innovators in the healthcare ecosystem. Inviting the healthcare and open source community to join in and build upon our platform. Leading to ever-smarter, adaptive, state-of-the-art models and applications, that will transform the healthcare landscape.

Always Open

Open Source

Generative AI – especially in healthcare – should be Always Open, to make sure it is safe, secure and utilizes the latest insights. That is why we invite anyone to review, customize and contribute to our code – creating a vibrant community of health AI innovators and users. We offer the flexibility to use multiple models, tools, and applications – which prevents vendor lock-in and offers you continuous access to state-of-the-art AI models that best fit your workflow.

Always Responsible

Quality & Compliance

Compliance is fundamental to our approach. We adhere to the most stringent and recent AI regulations, providing our customers flexibility without sacrificing regulatory standards. We aim to not only meet the highest industry standards but rather raise the bar, earning the trust of healthcare professionals worldwide. At the heart of our mission lies a relentless dedication to excellence and integrity, setting a new standard for responsible AI innovation in healthcare.

Always Safe

Trained by professionals

Using high-quality (licensed) health data, we are continuously training the most advanced Large Language Models (LLMs). As a next step, our models undergo Reinforcement Learning by Human Feedback, led exclusively by medical professionals. Our LLMs are then subjected to the same medical exams and certifications that health professionals need to pass. This way we ensure our LLMs are safe and trusted by the healthcare community.

Our launching application

Note-to-FHIR

In healthcare, there is a tremendous amount of knowledge captured in clinical notes and other types of unstructured text. Now for the first time in history, this unstructured data can be analyzed automatically, with unprecedented speed and accuracy. The first release of our Large Language Model, Note-to-FHIR, transforms unstructured data into a structured FHIR output, enabling seamless integration with your healthcare information systems.


In keeping with our Always Open and collaborative approach, we currently release a Beta version of Note-to-FHIR, to enable the (medical) community to review and extend our solution prior to the impending launch of the first commercial version.

State-Of-The-Art Accuracy

The HealthSage AI Note-to-FHIR LLM is State-of-the-Art, outperforming other and much larger models. Compared to GPT-4 it is 8% more accurate (93% vs 85%).

Resource Efficiency

Created by finetuning a fit-for-purpose open source LLM, our Note-to-FHIR model is magnitudes more resource-efficient than other commercial models – yielding cost savings and reducing environmental impact.

Transparent and Customizable

Our open source model allows for full customization and transparency. See what’s under the hood – no more black boxes – and stay free from vendor lock-in. Leverage community contributions, or even build your own extensions on top of our model.

Foundation for the future

The HealthSage AI Platform

The HealthSage AI platform will offer a rich selection of AI models, tools, and applications, both crafted by HealthSage AI and contributed by the broader community. In our secure environment, you can explore, try, and customize these models, before integrating them into your existing workflows and IT systems using our APIs.

By orchestrating multiple models and applications, HealthSage AI will help you to stay on top of the latest developments. We will make sure you always have access to state-of-the-art models and tools – leading to true clinical impact.

  • A trusted, secure and safe environment to build and share your solutions
  • A single API to integrate all available applications into your hospital information systems
  • State-of-the-art AI models that best suit your use case or workflow
  • Cloud agnostic, easy to deploy and fully customizable
Developer engagement

Build the future with us

The HealthSage AI platform is set to redefine the healthcare ecosystem, offering unparalleled possibilities for healthcare providers, organizations, and patients. If you – like us – champion the open source philosophy and are dedicated to transforming healthcare, then we invite you to leverage our platform as your foundation for the development of ground-breaking AI applications.

 

About HealthSage AI

We started HealthSage AI because we believe AI holds the key to solving many of the challenges healthcare faces today. Both our personal and professional experiences in healthcare sparked a desire to contribute to its future. With HealthSage AI, we aim to offer the healthcare and open source community a trusted and transparent foundation for numerous AI applications. These applications will help healthcare providers, organizations, and patients across the healthcare spectrum.

Our co-founders, Harm-Jan Wessels and Marcel Alberti, know what it takes to create solutions that make a genuine impact. At HealthSage AI, they have built a talented team with decades of invaluable experience in healthcare IT. Driven by creativity, curiosity and innovation, we are ready to shape the future of healthcare.

 

Updates

Want to keep up with all our latest news? Sign up now to stay in the loop.


Follow us for more updates: