BoltzGen
Type anything; we’ll convert to a safe ID when you run.
# API example will render hereShort label for this run. We will canonicalize it into a safe ID on submit.
Pick a preset protocol matching your modality (protein, peptide, small molecule, nanobody, antibody).
Total number of candidate designs to generate before filtering and ranking.
How many designs to keep after filtering. Think of this as your final shortlist size.
Number of recycle passes used during the design stage.
Number of sampling steps used during the design stage.
How many diffusion samples to generate in the design stage.
Number of recycle passes used during the affinity stage.
Number of sampling steps used during the affinity stage.
How many diffusion samples to generate in the affinity stage.
Number of recycle passes used during the fold stage.
Number of sampling steps used during the fold stage.
How many diffusion samples to generate in the fold stage.
Number of recycle passes used during the inverse fold stage.
Number of sampling steps used during the inverse fold stage.
How many diffusion samples to generate in the inverse fold stage.
Controls filtering aggressiveness in BoltzGen ranking.
Matches the CLI biased filtering switch.
Optional numeric threshold. Leave blank to disable.
Add at least one entity. Typical setup is a target structure + a designed protein or peptide. Ligands are optional.
This snippet submits the same payload the UI sends. Use workflow_name and boltzgen.name consistently.
Use “Sync” to include real selections; use “Reset” to return to a minimal example payload.
# API example will render here