DIAGNOSIS IS A MOMENT
HEALTH IS A TRAJECTORY
Enter Zenta
Two capabilities – One engine
Structuring
Turns clinical information into standardised, machine-ready data.
.svg)
Structuring turns clinical information into standardised, machine-ready data. Powered by LUWAK™, running on ARK™.
Connect.
ARK™ connects to existing hospital systems without changing workflows.
Structure.
LUWAK™, our on-site structuring engine, extracts clinically relevant information, adds context and standardises it on-premise.
Put to work.
Structured, interoperable data for AI, research and care.
Imaging
Turns medical images into quantitative information.
Imaging turns medical images into quantitative information. Two tools, one output: measurable data.
Segmented organs
Multi-organ segmentation of CT images: organs, heart, lungs, muscles, ribs, vertebrae, bones, body composition and more.
Segmented lesions
Fully automated tumour segmentation for FDG and PSMA PET scans, from PET input alone.
Output.
Segmentation masks plus quantitative metrics, such as volume, intensity and SUV values.
Runs best on ARK™
ARK™
The engine: Brings the processing to the data—on-site, not in an external cloud.
.svg)
ARK™ brings the processing to the data.
ARK™ is our GPU-powered engine: a high-performance workstation, scaled to the needs of our customers. Our tools run on ARK™, on-site, where health data is generated, not in cloud space.
Working across sites.
When several institutions collaborate, our de-identification tools on ARK™ prepare extracted data for a shared data lake. The lake and the transfer infrastructure remain with you and your partners.
Structured, quantitative foundations for Health Intelligence
Clinical Data Management
Structured information across patient records; from reports to images and handwritten notes
Research & Cohorts
De-identified, machine-ready data for studies, cohorts and collaboration across institutions.
AI Development
Consistent, quantitative data to train and validate AI models.
The road ahead
NOW
Structured clinical data and imaging
NEXT
A longitudinal view across provider networks
LONG TERM
Wearables and molecular biomarkers
FAQS
Answers to common questions about Zenta, ARK™, LUWAK™ and our imaging tools.
No. Zenta provides data infrastructure and analytics. Clinical interpretation and interventions remain with qualified healthcare teams.
Sensitive data does not have to be sent to an external cloud for processing. Institutions keep control over their data and infrastructure.
Yes. When institutions collaborate, our de-identification tools on ARK™ prepare extracted data for a shared data lake. The lake and the transfer infrastructure are provided by the participating institutions and their partners.
ARK™ is the engine our tools run on: a GPU-powered, high-performance workstation, installed on-site and scaled to each customer’s needs. It is required for LUWAK™ and recommended for our imaging tools.
Zenta provides tools that extract and structure clinically relevant information from medical records and medical images. They turn fragmented health data into standardised, machine-ready data for AI and research.
Processing happens on-site, where the data is generated. Data is structured and de-identified inside the institution before it is used anywhere else.
LUWAK™ currently supports German and English, and we have validated both. It extracts clinical parameters directly from each hospital’s original reports and handwritten notes, with no translation needed. Whatever the source language, the extracted data is mapped into the same structured format, so information from different hospitals can be compared and analysed together.
They can be processed too. LUWAK™ uses optical character recognition (OCR) to digitise handwritten notes, scanned forms and paper records. The recognised text then goes through the same extraction process as digital reports, so clinical parameters are captured and structured whatever format they started in. Information that would otherwise stay on paper becomes part of the hospital’s usable knowledge base.
Yes. Much of the most valuable clinical information sits in free-text documents such as radiology findings, discharge letters and pathology reports. LUWAK™ analyses these reports directly on-premise and extracts the key clinical parameters, such as diagnoses, measurements, lab values and medications, into a consistent, structured format. Narrative text becomes data that can be searched, compared and analysed, and it is de-identified before it ever leaves the hospital.
ARK™ is the hardware: a high-performance workstation installed on-premise at the hospital, which connects to existing systems and securely captures medical data at the source. LUWAK™ is the software that runs on ARK™: our intelligent structuring engine, which cleans and contextualises the data and turns it into structured, usable formats. Because LUWAK™ runs directly on ARK™, data is structured inside the hospital before it ever leaves. ARK™ collects and protects the data, and LUWAK™ turns it into knowledge.
No. Our imaging tools are not certified medical devices: they have no CE mark under the EU MDR and no FDA clearance. They are research software, not intended for clinical diagnosis or treatment decisions. Any results used in a clinical context should be reviewed by a qualified expert.
Our imaging tools run on ARK™ or on Windows, Mac and Linux, and need at least 16 GB RAM and Python 3.10+. An NVIDIA GPU is recommended but optional (CPU is slower). Apple Silicon runs natively via MPS.
Segmentation masks plus CSV files with quantitative metrics, such as volume, intensity and SUV values.
They segment organs and anatomy in CT images, and tumour lesions in FDG and PSMA PET scans.
Meet the team
Co-founder of the world’s first PET/CT, currently serving as a Professor of Medical Physics at the Medical University of Vienna, with 30 years of expertise in medical imaging.
Former Executive from GE Healthcare, who led numerous innovations in molecular imaging, bringing 34 years of experience in both industry and research within the field of medical imaging.
Serving as a Research Engineer at Zenta, David holds a PhD in Deep Learning for Medical Imaging, where he conducted research on Generative AI for Medical Image Harmonization and Synthetic Data Generation at the Medical University of Vienna.
Sebastian is a PhD student in Medical Physics at the Medical University of Vienna and a key contributor to the ENHANCE.PET initative. His work focuses on medical image registration, analysis, and processing, using both conventional methods and AI-assisted approaches.
career at zenta
Our team is united in one belief: that if we can build the right infrastructure for our medical data, we can make healthcare better for patients and care providers alike. If this resonates with you, we’d love to have you join us.
Built by Builders
We are building trust, knowledge, and a future for healthcare that is worthy of the people it serves.
Accountability
We hold ourselves accountable not only for the work we do, but the data that we are entrusted with.
Care
We are driven by our desire to make healthcare better and are looking for more people who share our vision.
Health Intelligence
From A to Zenta
Health intelligence comes from seeing the whole picture — not isolated symptoms, images or records. Zenta brings diverse health signals into a legible context, enabling others to connect the dots that make health understandable.
Book A Demo
Structuring, Imaging or both — see what Zenta can make legible in your setting.
GET IN TOUCH
Have questions or want to learn more? We’d love to hear from you!
