The problem
Modern labs run on sophisticated instruments. However, converting biological hypotheses into something robots can execute is still slow, manual work.
AI has made it easy to generate hypotheses and experimental designs, but without a verifier for biological ground truth, it isn’t easy to check these ideas in the physical world. That gap between the in silico and the wet lab is the bottleneck we’re working on.
Closing the loop from both sides of the wet lab.
We envision a future in which agents can design an experiment, run protocols in the wet lab, pull raw output off the instruments, generate structured data, and analyze information for biological meaning. Agents driving autonomous, closed-loop science provides our best answer for rapid basic research.
We work on both sides of the wet lab: the protocol going in, and the data coming out.
Agents read the protocol the way you wrote it. Project IOBring a protocolIn plain language: a PDF, a list of steps or a kit insert.Nothing moves until a person signs off. Project IOSimulate and approveVerified on digital twins before experiments get sent out. Instruments begin execution only after human approval.The data comes off as a spreadsheet. BioBase ReaderRun and read the dataThe protocol runs on your instruments, and the data comes off as CSV or Excel.Agents work through the results. BioBaseAnalyze the data off the instrumentsProtocol run data and analysis records are bundled together for audit and compliance purposes.
Labs are limited by time, people and capital.
Of the three, capital is usually the easiest to solve. People find money when it gives them a measurable chance at better discovery, or more of it.
The harder limit is the number of people in the lab. Once an agent can control the instruments and read the data off them, that limit matters much less.
If you don’t need someone babysitting one instrument, you don’t need someone babysitting six.
