ComplexesGPU15 credits

Chai-1 — Protein-Ligand Complex Prediction

Predict how drugs bind to protein targets at atomic resolution

Chai-1 predicts protein-ligand, protein-protein, and antibody-antigen complex structures at atomic resolution. An open-source alternative to AlphaFold3, Chai-1 achieves a 77% success rate on the PoseBusters benchmark and supports single-sequence mode without requiring multiple sequence alignments.

$0.15

per API call

credits per call

/v1/complexes/chai1

API endpoint

Features

Protein-ligand complex structure prediction

Single-sequence mode — no MSA required

Multi-chain complex support

Antibody-antigen interaction modeling

Per-residue confidence scores

GPU-accelerated on NVIDIA A100 80GB

Quick Start

Chai-1 — Python Examplepython

import requests
import time

API_KEY = "sk-sci-your-key-here"
url = "https://scirouter.ai/v1/complexes/chai1"

# Submit a complex prediction job
response = requests.post(url, json={
    "protein_sequence": "MKFLILLFNILCLFPVLAADNHGVS...",
    "ligand_smiles": "CC(=O)Oc1ccccc1C(=O)O"
}, headers={"Authorization": f"Bearer {API_KEY}"})

job = response.json()
print(f"Job ID: {job['job_id']}")

# Poll for results
while True:
    result = requests.get(f"{url}/{job['job_id']}",
        headers={"Authorization": f"Bearer {API_KEY}"})
    data = result.json()
    if data["status"] == "completed":
        print(f"Complex PDB: {len(data['pdb'])} bytes")
        print(f"Confidence: {data['confidence']:.3f}")
        break
    time.sleep(3)