AI-Assisted Structure Elucidation

Enter the spectra and I'll suggest the structure

Peak lists of ¹H NMR, ¹³C NMR and IR in; ranked candidate structures out — each one checked by predicting its spectra again.

¹H NMR

¹³C NMR

Mass spectrum (optional)

Peak of the molecular ion

What else do you know?

Molecular formula Molecular mass

Both are optional. A formula limits the candidates strictly; without one the mass is used.

Look for known structures with this formula in PubChem

Only the formula is sent to pubchem.ncbi.nlm.nih.gov, never your spectra. If nothing suitable is found, or PubChem cannot be reached, the AI model proposes the structures. A file data/library.smi (one SMILES per line, optional name) is searched too.

Already have a structure?

Structure to compare (SMILES)

Instead of searching, check one structure you already have: its spectra are predicted and matched against the data above. The formula and mass fields, if filled in, are checked against it.

Measurement conditions

NMR solvent Spectrometer frequency Frequency (MHz) IR measurement method Candidates

What should the analysis do?

Candidates, rankedEvery candidate's spectra are predicted again and compared with yours, then the candidates are ranked. Candidate listOnly the list of structures that fit the formula and the data, with the checks done on each. No spectra are predicted.

Examples:

The examples are approximate literature values, typed from general knowledge.

Candidates are AI hypotheses ranked by how well their predicted spectra resemble your data. This is not proof of a structure; confirm it with experiments.