High Content Screening Analysis
That's the state of the field today. We built something different.
Our technology extracts individual morphological features from fluorescence microscopy images; the same measurements you are already familiar with: cell and nuclear size, shape, texture, intensity distributions, and spatial relationships between organelles and compartments.
These features are the foundation of high content screening, toxicology assessment, spatial genomics, and digital pathology. We didn't invent new outputs. We made computing them 100× faster without sacrificing the interpretability your downstream analysis depends on.
Analysis runs on a custom-built, managed AI-inference server deployed on your site. No proprietary screening data is transmitted to the cloud. No vendor has access to your compounds, results, or imaging data.
For pharmaceutical research, this isn't a feature — it's a requirement. We designed the deployment model around it from the start.
CellProfiler is open source and widely used, but it was designed in an era before GPU computing and modern AI. Each image processes on the CPU, which is why a single plate takes 20 hours. That's not a bottleneck, it's a wall.
Commercial alternatives from Revvity, Molecular Devices, and Olympus face the same fundamental constraint, and are additionally tied to proprietary imaging hardware, limiting who can use them. None of them are architected to move faster.
We were built from the ground up around GPU kernel engineering and AI inference, not retrofitted onto a CPU pipeline. The speed improvement is fundamental.
Our minimum viable product for high content screening assays is available now. We are signing first paid pilot customers and beginning beta testing. A peer-reviewed benchmark publication is underway.
We are a small team moving quickly. If you are running HCS campaigns and spending meaningful time waiting on analysis, we want to hear from you.
We're working with a small group of researchers during our beta. Tell us about your work.