live · real Hartree–Fock · runs in this tab

Screen a molecule library,
live, in your browser.

Every candidate below gets a real quantum-chemistry calculation — a full Hartree–Fock SCF in WebAssembly, no server, no install — and we rank them by HOMO–LUMO gap, the property that sets a dye's color, a molecule's reactivity, its stability. Watch the board sort itself as the numbers land.

The library: isoelectronic aza-chains (carbons swapped for nitrogens on a conjugated backbone). Same electron count, so the gap differences are purely where the nitrogen sits. The question the screen answers blind: do the top chromophores share the azo (–N=N–) motif behind most commercial dyes?

~30–60 s · 12 real SCF runs · STO-3G · everything computed on your machine
0/0
molecules computed
gap range (eV)
azo enrichment in top hits
wall time
absorption-shift spectrum
◀ smaller gap · red-shifted larger gap · blue-shifted ▶

Each dot is a molecule, placed by its HOMO–LUMO gap. Glowing pink = azo (N=N) present. Color is the relative shift across this library — HF/STO-3G overestimates the absolute wavelength (real dyes absorb in the visible; minimal-basis HF puts these in the UV), but the trend it ranks on is the robust, validated part.

candidates — ranked smallest gap first
what the screen found
AZO-MOTIF ENRICHMENT
azo in top 15
azo across whole library

Run the screen — if azo is more common at the top than overall, the descriptor is tracking real dye chemistry, not noise.

LIVE COMPUTE LOG · proof it's real
idle — press “Run the screen”.
each line is one real SCF: iterations, convergence,
engine (wasm/gpu), and wall time.

Every number here is produced by the same engine that's cross-checked against PySCF / FCI in CI (HF within ~50 µHa). This is triage-grade screening, not a proven hit list — minimal-basis HF overestimates gaps and can break down into artifacts (flagged above). The point isn't to beat PySCF; it's that a real wavefunction calculation now ships as a URL. See also the 4D hyperscope · distributed swarm · SI report.