ProteinsGPU15 credits

BoltzGen — Universal Binder Designer

Design protein binders against any target with predicted binding affinity

BoltzGen (MIT Jameel Clinic, 2026) is the successor to RFdiffusion for universal binder design. Given just a target sequence (no structure needed), BoltzGen generates candidate binder sequences with predicted ΔG binding affinity and confidence scores. SciRouter exposes BoltzGen via a single API call.

$0.15

per API call

credits per call

/v1/design/binder/boltzgen

API endpoint

Features

Sequence-only binder design (no target PDB needed)

Predicted ΔG binding affinity per design (kcal/mol)

Interface residue identification

Confidence score per binder

Async job pattern with result polling

Complements RFdiffusion for backbone + sequence parity

Quick Start

BoltzGen — Python Examplepython

import requests

API_KEY = "sk-sci-your-key-here"
url = "https://scirouter.ai/v1/design/binder/boltzgen"

# PD-L1 extracellular domain sequence
pdl1 = "FTVTVPKDLYVVEYGSNMTIECKFPVEKQLDLAALIVYWEMEDKNIIQFVHGEEDLKVQHSS..."

resp = requests.post(
    url,
    json={
        "target_sequence": pdl1,
        "hotspots": [42, 67, 89],
        "length_min": 65,
        "length_max": 85,
        "num_designs": 5,
    },
    headers={"Authorization": f"Bearer {API_KEY}"}
)
job = resp.json()["data"]
for d in job["result"]["designs"]:
    print(f"  {d['design_id']}: ΔG={d['predicted_delta_g_kcal_mol']} conf={d['confidence']}")