Our vision
Cancer cases are rising and diagnostic requirements are growing more complex, faster than the number of pathologists. We build the foundation models and products that let a pathologist measure rather than estimate, and keep the decision.
Demand is outpacing pathologists
Modern medicine faces a critical challenge: a rising number of cancer cases and increasingly complex diagnostic requirements are outpacing the number of available pathologists. Laboratories are under pressure, delays grow, and specialist workload rises.
Manual analysis remains the gold standard, but it is time-intensive and varies between observers.
AI that supports the pathologist
We believe AI is the key to empowering pathologists. By handling tasks like cell counting and tissue segmentation, our models act as a tireless digital assistant, automating the routine and returning quantitative insight.
Our goal is not to replace pathologists but to support them with intuitive, reliable and seamlessly integrated tools, developed together with medical institutions so that they stay clinically relevant.
Principles
Evidence first
We publish our methods and validation before we make claims about them.
Clinical co-development
Every product starts as a question from a pathologist and is validated on their cases.
The pathologist decides
Overlays are there to be checked. The model provides the count; the diagnosis remains with the pathologist.
A UCL spinout
Octopath grew out of a UKRI Future Leaders Fellowship at UCL in AI for cancer diagnosis, and was launched as a research platform in 2024. Kraken, our H&E foundation model trained on 14.1 million labelled cells, was released in June 2026.
Read the news →Work with us
We are hiring scientists and engineers, and we partner with clinical and research institutions.