A GPS for Health, built as open infrastructure.

Cytognosis builds open-science AI infrastructure that turns molecular, cellular, physiological, behavioral, and clinical research signals into interpretable health-state coordinates. The goal is to help researchers understand disease as a trajectory, long before a clinical label can explain it. This work is research infrastructure, and it is non-diagnostic by design.

Why a new map is needed

Reactive healthcare has three structural blindspots.

Modern medicine is good at naming conditions once symptoms appear. It is far weaker at seeing the biology underneath, the years before a label fits, and the differences between people. Cytognosis builds the shared AI layer that helps researchers address all three.

Mechanistic: symptoms are not causes
We group people by symptom clusters, not by root biology. Depression has several molecular subtypes, and autoimmune conditions often take years and many specialists to identify. Shared mechanisms that cut across diagnoses stay hidden.
Temporal: meaningful change starts early
Many disease processes begin as small deviations from a personal baseline years before symptoms. Alzheimer's pathology, for example, develops well before memory loss. A label that arrives late describes the destination, not the trajectory.
Complexity: people are not averages
Continuous glucose monitoring showed that identical meals produce very different responses across people (Zeevi et al., Cell, 2015). Binary categories collapse that biological variation into a population average and lose the individual.

These are not three separate problems. They are one structural gap: the absence of a shared system that maps health at its biological roots, follows it over time, and keeps every step interpretable and uncertainty-aware.

The platform

GPS for Health: a map, a sensor, and a navigator.

A GPS needs a map, a way to take a reading, and a navigator to make sense of it. Cytoverse, Cytoscope, and Cytonome are designed to build, measure, and explain health-state coordinates across biological, behavioral, and clinical scales.

The map
Cytoversethe map

An AI health coordinate system. Multimodal foundation models and ontology-guided embeddings align biological and behavioral signals into continuous, interpretable coordinates, in place of binary labels.

The sensor
Cytoscopethe sensor

A sensing and measurement layer. It tracks movement through the map over time using research-grade sensors, biomarkers, wearables, and validated instruments.

The navigator
Cytonomethe navigator

A privacy-aware AI interface. It turns coordinates into human-reviewable explanations with provenance, privacy boundaries, and explicit uncertainty.

Underneath sits the Helix Model, a multimodal biological foundation model that connects genomic, single-cell, connectomic, phenotype, and longitudinal signals into one research representation layer.

Foundation thesis

Disease is a trajectory, not a label.

Cytognosis models health as a dynamic position in biological state space. A meaningful shift may begin as a small deviation from a personal baseline across interpretable axes, long before a clinical label can explain it. Our work is to make those trajectories measurable, traceable, and uncertainty-aware, so the research community can study them and challenge them.

Why a nonprofit foundation

Public-good infrastructure, built in the open.

A shared map of human health is infrastructure, not a single product. Building it well needs sustained, foundational work that neither academia nor industry is structured to carry alone. A nonprofit foundation can hold that work for the public good.

The gap a foundation fills

Academia excels at foundational discovery but is not built for industrial-scale, long-horizon engineering and validation. Industry is structured around shorter-term products. A focused research organization can do the patient, coordinated work in between, and treat the result as a public utility that other groups build on.

Open by default

We build reusable models, schemas, benchmarks, validation reports, and documentation that research groups, patient communities, and institutions can evaluate and extend. Models and tools are released under open licenses when safe and legally permitted, with bias audits, performance bounds, and negative results documented so the field can inspect the work.

Take part

Help build the map.

For researchers, clinicians, engineers, patient communities, and funders working toward reproducible, equitable, and interpretable health-state AI.

Cytognosis is research infrastructure and does not provide medical diagnosis, treatment, or clinical decision-making.