Spectra Predictor
Enter a structure and get its 1H NMR, 13C NMR and IR spectra — shifts, coupling constants, multiplets and band assignments. Upload a measured spectrum and the analyser tells you what the structure explains and what it does not.
Input guidelines & notation+
Structures are entered as SMILES, the standard line notation for molecules.
CCO— chain of atoms; hydrogens are implicit. Ethanol.CC(=O)O— branches in parentheses,=for a double bond,#for a triple bond. Acetic acid.c1ccccc1— lowercase letters mean aromatic. Benzene. Kekulé formC1=CC=CC=C1works too.C/C=C/C—/and\set the double-bond geometry, which decides whether 3J is 10 or 16 Hz.[nH],[N+],[O-]— brackets carry explicit hydrogens, charges and isotopes.CCO.CC— a full stop separates disconnected components.
Stereocentres are parsed but not used: 1H shifts are predicted for an achiral average, so diastereotopic protons are reported as one signal.
Structure summary
Scroll to zoom · drag to select a region · shift-drag to pan · double-click to reset
Compare a measured spectrum
Drop a .jdx / .dx JCAMP-DX file, a two-column CSV, or a plain
peak list. Whatever the prediction cannot account for is matched against the
trace-impurity table for the selected solvent.
No file to hand? — a synthetic 1H spectrum of ethyl acetate spiked with acetone, water and dichloromethane.
Identify a structure from a spectrum
The other direction: give the analyser a measured peak list and it works out what the data already settles — how many carbons of each type, how many protons, which fragments are forced — then builds structures that fit and ranks them by predicted spectrum.
Click two points on the chemical-shift axis, then type what they read. Everything else is measured from the image.