First Commercial Platform to Predict Albumin Binding Site Specificity
Spirographic AI Becomes First Commercial Platform to Predict Albumin Binding Site Specificity — Achieving 86.4% Accuracy Across 198 Drugs
Every existing commercial tool predicts only binary bound/unbound status. spirographic AI predicts which site — Site I, Site II, or Subdomain IB — with 93.8% sensitivity for high binders, from SMILES input alone.
CORVALLIS, Ore., April 23, 2026 /PRNewswire/ — Spirographic AI, LLC today announced the commercial availability of its Albumin Binding Prediction Engine — the first and only platform to predict human serum albumin binding site specificity as a standard output. Every major ADMET prediction platform currently on the market, including ADMETlab, pkCSM, SwissADME, admetSAR, DruMAP, and ANDROMEDA, provides only binary or fractional plasma protein binding values. None predict which albumin binding site a drug occupies. Spirographic AI does — validated across 198 drugs with 86.4% overall accuracy and 93.8% sensitivity for clinically significant high binders.
Binding site identity is not a pharmacological footnote. Site I and Site II binders compete for albumin occupancy differently, interact with co-administered drugs differently, and behave differently under conditions of hypoalbuminemia, renal impairment, and glycation. Knowing that a drug is highly protein-bound is useful. Knowing where it binds is actionable — for formulation, for interaction screening, and for understanding displacement risk in polypharmacy patients.
“The entire ADMET prediction market has treated albumin binding as a yes/no question for decades,” said Brandy Stoffel, MSN, RN founder of Spirographic AI. “The question that actually matters in the clinic is where. We answer that — at 86.4% accuracy, from a SMILES string, fully automated. That capability does not currently exist anywhere else commercially.”
The Albumin Binding Prediction Engine is one module within the broader Spirographic AI platform, which delivers comprehensive pharmacokinetic predictions from SMILES input alone — including CYP450 metabolite generation (94.4% accuracy across 56 drugs on two independent benchmarks), blood-brain barrier, hepatic, renal, placental, pulmonary, retinal, and breast milk transporter substrates. The platform is available on a contract basis with zero data retention and is patent pending.
About Spirographic AI
Spirographic AI, LLC is an Oregon-based computational pharmaceutical transporter substrate and metabolite prediction platform founded by Brandy Stoffel, MSN, RN. The platform is patent pending and operates under a zero data retention policy.
Media Contact
Brandy Stoffel, MSN, RN
Founder, Spirographic AI, LLC
https://spirographicai.com/
503.352.4969