kanad
Quantum-Native Digital Laboratory
A web-based quantum chemistry platform that makes real quantum computation accessible through an intuitive laboratory interface. Build molecules, run quantum solvers, and analyse twenty-plus molecular properties — no quantum programming required.
The Bridge Between Chemistry and Quantum
Kanad is a web-based quantum chemistry platform built on the Kanad governance-driven quantum chemistry framework. It provides scientists and researchers with a complete workflow — from molecular design to quantum circuit execution to property analysis — without requiring expertise in quantum programming or circuit construction.
The platform runs actual quantum calculations using solvers like PhysicsVQE, which achieves chemical accuracy (< 1.6 mHa error) on molecules including H₂, HeH⁺, LiH, and H₂O. Validated on IBM Quantum hardware with 0.00 mHa error using HybridSubspaceVQE. Every data point in the interface comes from real Kanad computation modules — nothing is mocked or approximated.
The interface is designed around a dashboard + labs model — Schrödinger Lab for quantum calculations, Prigogine Lab for dynamics and reactions — with a premium research aesthetic: sharp edges, serif typography, warm gold accents. A professional research instrument, not a consumer application.
"Nature isn't classical, dammit!
— Richard Feynman
and if you want to make
a simulation of nature,
you'd better make it
quantum mechanical."
Chemical Bonding Governance Protocol
A novel framework that governs molecular simulation through bond-type-specific rules
The Physics of Bonding Dictates the Quantum Circuit
Current quantum chemistry frameworks apply the same generic circuit to every molecule — whether covalent H₂, ionic LiH, or metallic iron. This wastes quantum resources: 200+ evaluations for a two-electron system.
Kanad automatically detects bonding character — covalent, ionic, or metallic — from electronegativity differences and constrains quantum circuits to physically relevant operations through governance protocols. This achieves sub-milliHartree accuracy in 17 evaluations for H₂ — a 49x speedup over generic approaches.
"The physics of bonding should dictate the quantum representation, not the other way around."
Technical Architecture
Schrödinger Lab
Build molecules visually with 3D viewer, configure quantum solvers (PhysicsVQE, SQD, HardwareVQE), run on local or cloud backends (IBM Quantum, BlueQubit, IonQ), and analyse 20+ properties — from UV-Vis spectra to ADME drug discovery metrics.
PhysicsVQE — 49x Faster
MP2-guided excitation ranking + sequential 1D optimisation + frozen core. Chemical accuracy (< 1.6 mHa) in ~20 evaluations for diatomics, versus 200+ for generic ansatze. Validated on IBM Quantum hardware.
Prigogine Lab — Dynamics
Born-Oppenheimer molecular dynamics with quantum VQE forces. Configurable integrators (Velocity Verlet, Leapfrog), thermostats (Berendsen, Nosé-Hoover, Langevin), and chemical reaction PES scans with transition state search.
Cloud & Environment
Run on statevector, Aer noise-simulator, IBM Quantum hardware, BlueQubit GPU (36 qubits), or IonQ. Temperature, solvent (40+ solvents), pressure, and pH modulation for realistic conditions.
Research Domains
From molecular discovery to industrial optimization
Drug Discovery
Binding affinity estimation, ADME prediction (logP, Lipinski Rule of Five, BBB penetration), conformational sampling, and HOMO-LUMO gap analysis for pharmaceutical research.
Materials Science
Electronic structure screening at quantum fidelity. Metallic governance for catalysts and electrodes. Density of states, band analysis, and energy caching for batch computation.
Chemistry Research
Reaction pathway analysis, transition state search, thermochemistry (enthalpy, Gibbs free energy, entropy), PES scans, and vibrational frequency calculations.
Biochemistry
Active site modelling, spectroscopic properties (UV-Vis, NMR, Raman, IR), dipole moments, polarisability tensors, and isotope effects for biological systems.
Quantum Environment Is All You Need
Reinforcement Learning agents that learn chemistry from real quantum eigensolver computations
Research Paper
We present Gymnasium environments where every reward signal is a real VQE quantum computation. Energy-gradient rewards produce genuine learning — 71% to 94% near-equilibrium rate, 5.4x error reduction — while exploration rewards show no learning. Built on Kanad governance (49x faster), with 93% cache hit rate at 4,096 timesteps.
RL Adaptor — Coming Soon
Four Gymnasium-compatible environments where every energy is a real VQE computation. Agents learn equilibrium distances, PES shapes, which atoms bond, and where barriers are. Standard Gymnasium API — works with any RL library. The adaptor for training RL models on quantum ground truth will be launched very soon.
MVP Launch — 24 April 2026
Kanad enters its MVP phase with core quantum chemistry workflows
MVP Launch
Schrödinger Lab with PhysicsVQE solver, molecule builder, and core property analysis on local and cloud backends
Beta — Full Labs
Prigogine Lab (dynamics & reactions), campaigns, reports, and expanded cloud hardware access
Production
Enterprise features, API access, and RL environment adaptor for training agents on quantum ground truth