A thousand seeds, one binder that binds.
Two high-schoolers from Georgia, three faculty letters of intent, and one workflow we wrote because nothing else got us a binder before the wet-lab window closed.
The name
Japanese legend: fold a thousand paper cranes and you are granted one wish. In our pipeline a thousand seeded designs yield, on a good week, one binder that actually binds. The math rhymes. So does the patience.
Origami isn’t a metaphor either. Erik Demaine (MIT) proved paper-folding is computationally equivalent to protein folding — both are flat sheets searching for a single low-energy 3D shape. The crane is also an East Asian symbol of healing. Therapeutics are the end state.
“A thousand folds for one wish. A thousand seeds for one binder. The patience is the product.”
Designed, not discovered.
We spent a full iGEM season running ProteinMPNN, RFdiffusion, and BindCraft against the same target — by hand, one seed at a time, on three different clusters with three different broken Conda envs. We lost weekends to CUDA mismatches before we shipped a single binder that bound. The tools were all there. The workflow wasn’t.
Paper Crane is that workflow. Every major tool, behind one API, with one cost meter and one pre-registered scoring pipeline. You bring a target. We hand you ranked binders. Free for academics, $299/mo for labs, custom for pharma. No Conda. No bash for-loops. No three-day debugging sessions over a CUDA version mismatch.
The next generation of therapeutics will be designed, not discovered. The platform underneath should be as elegant as the proteins it builds. That’s the only thing we’re trying to make true.
Paper Crane opens to a small cohort of academic labs on 2026-04-30. If your wet-lab window is closing, write to hello@papercrane.bio — we’ll get you a seat.
iGEM co-captain, 2025 international winner. Wrote the original Mosaic pipeline — Protenix2025 → Modal/Vast.ai batch → Boltz2 → ipSAE/DockQ filtering — and ran the 50-seed campaigns it now schedules. Builds in public.
iGEM co-captain, 2025 international winner. Wet-lab lead. The reason any of our designs actually express, fold, and bind on the bench.
LOI on file. Public listing pending permission.
LOI on file · computational protein design and de novo binders. Public listing pending permission.
LOI on file · structural biology and ML for biomolecules. Public listing pending permission.
Reserved for an operator who has shipped a therapeutic or a platform. Intros welcome.