About
The models are open. The access isn't.
Corollary Labs exists to close one specific gap: between the biological AI models that already exist in the open, and the researchers who have the questions but not the cluster, not the toolchain, and not three days to spend on either.
- 542
- Models and datasets
- 191
- Ready to run
- 6
- Free forever
indexed and searchable
input to output, no setup
no account, no card
Why the name
One thing follows another.
A corollary is what follows necessarily from what came before. That is the shape of the work: given a structure, a binding prediction follows. Given a binding prediction, a design follows. The chain is obvious to anyone who knows the field — and almost impossible to actually run.
The gap between reading a paper about a folding model and folding your own protein of interest is not measured in minutes. It is measured in days of dependency conflicts, if the hardware exists at all. Most open models are hosted nowhere. The ones you can download won't fit on a laptop. The databases are scattered across niche environments with their own schemas and quirks. And when you finally have all the pieces, nothing tells you which to chain together, in what order, or how to read the result.
So we built the layer that does. Every model reachable the same way, whatever it was written in. Something that reads your question, picks the right model or sequence of models, and shows you why it picked them. The public databases already connected, already searchable. And an agent that runs the chain and grounds every claim it makes in something a tool actually returned.
The goal is simple to state and hard to reach: make biotech AI as accessible as sending a prompt to a chatbot.
The team
Built by people who kept hitting this wall.
Software engineering, computational biology, theoretical physics and food chemistry — spanning both sides of the gap between a model and the biologist who needs it.
Gabriele Battimelli
Software Engineering · AI/ML · Product
Over 10 years of software development. Designs and builds AI/ML pipelines and software infrastructure at Kernel Science. Advisor to European Commission President Ursula von der Leyen on digital policy. Science communicator on TV and radio.
Simone Anzà, PhD
Computational Biology · Multi-omics
Venture Fellow 2026, WashU School of Medicine. Former postdoc at SegataLab (CIBIO, University of Trento). Currently working on the epigenetic effects of stress, metagenomics, the microbiome-gut-brain axis and neurodevelopment at WashU.
Fabio Anzà, PhD
Theoretical Physics · Quantum
PhD, University of Oxford. Research Fellow at UC Davis; Research Assistant Professor at the University of Washington's InQubator for Quantum Simulations. Now Professor at UMBC.
Bruna Anzà
Food Chemistry · Sustainable Biotech
PhD candidate in Chemical Engineering, Politecnico di Torino. Former Team Leader at Alpha-Protein GmbH. Scientific Consultant, The Good Food Institute Europe. Vice-President, Cell Ag Italy.
Get in touch
Tell us what you're trying to run.
Whether it's a model we don't have yet or a deployment your legal team needs to approve, one of us reads every message.
Researchers and labs
A model missing from the library, a result that looks wrong, a workflow that should be one click and isn't. These are the messages that change what we build next.
Biotech, CRO and pharma
Private deployment inside your own network, data residency in your region, your own model registry, and volume pricing. Research data is never used for training.
Will you join us in making biotech AI as accessible as sending a prompt?
Free to start. Nothing to install.