Start a small molecule de novo design run

$ boltz-api small-molecule:design start

POST/compute/v1/small-molecule/design

Create a new design run that generates novel small molecule candidates for a protein target

ParametersExpand Collapse

--num-molecules: number

Number of molecules to generate. Must be between 10 and 1,000,000.

minimum10

maximum1000000

--target: object { entities, bonds, constraints, 2 more }

Target protein with binding pocket for small molecule design or screening

--chemical-space: optional "enamine_real"

Chemical space to constrain generated molecules. Currently only 'enamine_real' (Enamine REAL chemical space) is supported. Additional options may be added in the future.

--idempotency-key: optional string

Client-provided key to prevent duplicate submissions on retries

maxLength255

--molecule-filters: optional object { boltz_smarts_catalog_filter_level, custom_filters }

Molecule filtering configuration. Controls both Boltz built-in SMARTS filtering and custom filters.

--workspace-id: optional string

Target workspace ID (admin keys only; ignored for workspace keys)

ReturnsExpand Collapse

SmallMoleculeDesignStartResponse: object { id, completed_at, created_at, 12 more }

A small molecule design engine run that generates novel molecules

id: string

Unique SmDesignRun identifier

completed_at: string

formatdate-time

created_at: string

formatdate-time

data_deleted_at: string

When the input, output, and result data was permanently deleted. Null if data has not been deleted.

formatdate-time

engine: "boltz-sm-design"

Engine used for small molecule design

engine_version: string

Engine version used for small molecule design

error: object { code, message, details }

code: string

Machine-readable error code

message: string

Human-readable error message

details: optional unknown

Additional field-level error details keyed by input path, when available.

input: object { num_molecules, target, chemical_space, 3 more }

Pipeline input (null if data deleted)

num_molecules: number

Number of molecules to generate. Must be between 10 and 1,000,000.

minimum10

maximum1000000

target: object { entities, bonds, constraints, 2 more }

Target protein with binding pocket for small molecule design or screening

entities: array of object { chain_ids, type, value, 2 more }

Protein entities defining the target structure. Each entity represents a protein chain.

chain_ids: array of string

Chain IDs for this entity

type: "protein"

value: string

Amino acid sequence (one-letter codes)

cyclic: optional boolean

Whether the sequence is cyclic

modifications: optional array of object { residue_index, type, value } or object { residue_index, type, value }

Post-translational modifications. Optional; defaults to an empty list when omitted.

Accepts one of the following:

CCDModificationResponse: object { residue_index, type, value }

residue_index: number

0-based index of the residue to modify

minimum0

type: "ccd"

value: string

CCD code from RCSB PDB (e.g. 'MSE' for selenomethionine, 'SEP' for phosphoserine)

SmilesModificationResponse: object { residue_index, type, value }

residue_index: number

0-based index of the residue to modify

minimum0

type: "smiles"

value: string

SMILES string for the modification

bonds: optional array of object { atom1, atom2 }

Covalent bond constraints between atoms in the target complex. Atom-level ligand references currently support ligand_ccd only; ligand_smiles is unsupported.

atom1: object { atom_name, chain_id, type } or object { atom_name, chain_id, residue_index, type }

Ligand atom reference. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

Accepts one of the following:

LigandAtomResponse: object { atom_name, chain_id, type }

Ligand atom reference. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

atom_name: string

Standardized atom name (verifiable in CIF file on RCSB). Atom-level references to ligand_smiles entities are currently unsupported; use ligand_ccd instead.

chain_id: string

Chain ID containing the atom

type: "ligand_atom"

PolymerAtomResponse: object { atom_name, chain_id, residue_index, type }

atom_name: string

Standardized atom name (verifiable in CIF file on RCSB)

chain_id: string

Chain ID containing the atom

residue_index: number

0-based residue index

minimum0

type: "polymer_atom"

atom2: object { atom_name, chain_id, type } or object { atom_name, chain_id, residue_index, type }

Ligand atom reference. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

Accepts one of the following:

LigandAtomResponse: object { atom_name, chain_id, type }

Ligand atom reference. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

atom_name: string

Standardized atom name (verifiable in CIF file on RCSB). Atom-level references to ligand_smiles entities are currently unsupported; use ligand_ccd instead.

chain_id: string

Chain ID containing the atom

type: "ligand_atom"

PolymerAtomResponse: object { atom_name, chain_id, residue_index, type }

atom_name: string

Standardized atom name (verifiable in CIF file on RCSB)

chain_id: string

Chain ID containing the atom

residue_index: number

0-based residue index

minimum0

type: "polymer_atom"

constraints: optional array of object { binder_chain_id, contact_residues, max_distance_angstrom, 2 more } or object { max_distance_angstrom, token1, token2, 2 more }

Structural constraints (pocket and contact). Atom-level ligand references currently support ligand_ccd only; ligand_smiles is unsupported.

Accepts one of the following:

PocketConstraintResponse: object { binder_chain_id, contact_residues, max_distance_angstrom, 2 more }

Constrains the binder to interact with specific pocket residues on the target.

binder_chain_id: string

Chain ID of the binder molecule

contact_residues: map[array of number]

Binding pocket residues keyed by chain ID. Each key is a chain ID (e.g. "A") and the value is an array of 0-indexed residue indices that define the pocket on that chain.

max_distance_angstrom: number

Maximum allowed distance in Angstroms between binder and pocket residues. Typical range: 4-8 A.

type: "pocket"

force: optional boolean

Whether to force the constraint

ContactConstraintResponse: object { max_distance_angstrom, token1, token2, 2 more }

Contact constraint between two tokens. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

max_distance_angstrom: number

Maximum distance in Angstroms

token1: object { chain_id, residue_index, type } or object { atom_name, chain_id, type }

Ligand contact token. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

Accepts one of the following:

PolymerContactTokenResponse: object { chain_id, residue_index, type }

chain_id: string

Chain ID

residue_index: number

0-based residue index

minimum0

type: "polymer_contact"

LigandContactTokenResponse: object { atom_name, chain_id, type }

Ligand contact token. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

atom_name: string

Atom name. Atom-level references to ligand_smiles entities are currently unsupported; use ligand_ccd instead.

chain_id: string

Chain ID

type: "ligand_contact"

token2: object { chain_id, residue_index, type } or object { atom_name, chain_id, type }

Ligand contact token. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

Accepts one of the following:

PolymerContactTokenResponse: object { chain_id, residue_index, type }

chain_id: string

Chain ID

residue_index: number

0-based residue index

minimum0

type: "polymer_contact"

LigandContactTokenResponse: object { atom_name, chain_id, type }

Ligand contact token. Atom-level ligand references currently support ligand_ccd entities only; ligand_smiles is unsupported.

atom_name: string

Atom name. Atom-level references to ligand_smiles entities are currently unsupported; use ligand_ccd instead.

chain_id: string

Chain ID

type: "ligand_contact"

type: "contact"

force: optional boolean

Whether to force the constraint

pocket_residues: optional map[array of number]

Binding pocket residues, keyed by chain ID. Each key is a chain ID (e.g. "A") and the value is an array of 0-indexed residue indices that define the binding pocket on that chain. When provided, these residues guide pocket extraction and add a derived pocket constraint during affinity predictions. That derived constraint remains separate from any explicit pocket constraints in target.constraints. When omitted, the model auto-detects the pocket.

reference_ligands: optional array of string

Reference ligands as SMILES strings that help the model identify the binding pocket. When omitted, a set of drug-like default ligands is used for pocket detection.

chemical_space: optional "enamine_real"

Chemical space to constrain generated molecules. Currently only 'enamine_real' (Enamine REAL chemical space) is supported. Additional options may be added in the future.

idempotency_key: optional string

Client-provided key to prevent duplicate submissions on retries

maxLength255

molecule_filters: optional object { boltz_smarts_catalog_filter_level, custom_filters }

Molecule filtering configuration. Controls both Boltz built-in SMARTS filtering and custom filters.

boltz_smarts_catalog_filter_level: optional "recommended" or "extra" or "aggressive" or "disabled"

Controls the stringency of Boltz's built-in SMARTS structural alert filtering, which removes molecules matching known problematic substructures. 'recommended' (default): applies a curated set of alerts balancing safety and hit rate. 'extra': adds additional alerts beyond the recommended set for stricter filtering. 'aggressive': applies the most comprehensive alert set — may reject viable molecules. 'disabled': turns off Boltz SMARTS filtering entirely; only custom_filters will be applied.

Accepts one of the following:

"recommended"

"extra"

"aggressive"

"disabled"

custom_filters: optional array of object { max_hba, max_hbd, max_logp, 3 more } or object { type, fraction_csp3, mol_logp, 8 more } or object { patterns, type } or 2 more

Custom filters to apply. Molecules must pass all filters (AND logic).

Accepts one of the following:

LipinskiFilterResponse: object { max_hba, max_hbd, max_logp, 3 more }

Lipinski's Rule of Five filter. Rejects molecules that violate drug-likeness criteria based on molecular weight, LogP, hydrogen bond donors, and hydrogen bond acceptors.

max_hba: number

Maximum number of hydrogen bond acceptors. Lipinski threshold: 10

max_hbd: number

Maximum number of hydrogen bond donors. Lipinski threshold: 5

max_logp: number

Maximum LogP. Lipinski threshold: 5

max_mw: number

Maximum molecular weight (Da). Lipinski threshold: 500

type: "lipinski_filter"

allow_single_violation: optional boolean

If true, one rule violation is allowed (classic Rule of Five). Defaults to false (all rules must pass).

RdkitDescriptorFilterResponse: object { type, fraction_csp3, mol_logp, 8 more }

Filter molecules by RDKit molecular descriptors. Each descriptor is constrained to a min/max range. Only descriptors you provide are checked — omitted descriptors are unconstrained.

type: "rdkit_descriptor_filter"

fraction_csp3: optional object { max, min }

Min/max range constraint for an RDKit molecular descriptor

max: optional number

Maximum allowed value (inclusive)

min: optional number