Build binders. Ship them.
Paper Crane orchestrates twelve open-source models behind one authenticated API. These docs cover the platform end-to-end — from your first hotspot pick to a wet-lab-ready ranked list.
Quickstart
Authenticate, post a target, watch the campaign fold. The same call works against every model in the catalog.
curl https://api.papercrane.bio/v1/jobs \
-H "Authorization: Bearer $PAPERCRANE_KEY" \
-H "Content-Type: application/json" \
-d '{
"tool": "bindcraft",
"target_pdb": "2LGV",
"hotspots": ["A47","A52","A55"],
"n_designs": 100
}'Tool catalog
Twelve models, one runtime. Pick by stage of your pipeline.
Generation
- RFdiffusion/docs/tools/rfdiffusion
Diffusion-based de novo backbones with hotspot conditioning.
- BindCraft/docs/tools/bindcraft
End-to-end binder hallucination with built-in AF2 validation.
- BoltzGen/docs/tools/boltzgen
All-atom diffusion design — side-chain aware from step zero.
- Chroma/docs/tools/chroma
Programmable structure generation with conditional sampling.
Sequence
Structure
- AlphaFold 3/docs/tools/alphafold3
Multi-modal complex prediction across protein, NA, ligand.
- Boltz-1/docs/tools/boltz1
Open biomolecular structure prediction at AF2-class accuracy.
- Boltz-2/docs/tools/boltz2
Affinity-aware successor with improved interface scoring.
- Chai-1/docs/tools/chai1
Foundation model for protein, ligand, and nucleic-acid complexes.