DockingGPU10 credits

DiffDock — AI Molecular Docking

Diffusion-based molecular docking with confidence-ranked poses

DiffDock uses a diffusion generative model to predict protein-ligand binding poses without predefined search boxes. It generates multiple binding poses ranked by confidence score, outperforming traditional docking methods on many benchmarks.

Try It Now

No Login Required

Ligand SMILES

Examples:

Demo shows ligand properties. Full DiffDock docking requires a protein target + GPU.

$0.10

per API call

credits per call

/v1/docking/diffdock

API endpoint

Features

No predefined search box required

Confidence-ranked binding poses

GPU-accelerated inference

Supports PDB protein + SMILES ligand input

Up to 40 poses per prediction

Diffusion-based generative model

Quick Start

DiffDock v1.1 — Python Examplepython

import requests

API_KEY = "sk-sci-your-key-here"
url = "https://scirouter.ai/v1/docking/diffdock"

response = requests.post(url, json={
    "protein_pdb": open("target.pdb").read(),
    "ligand_smiles": "CC(=O)Oc1ccccc1C(=O)O",  # Aspirin
    "num_poses": 10
}, headers={"Authorization": f"Bearer {API_KEY}"})

job = response.json()
# Poll for results with job_id...