Loading VerChem
Preparing your chemistry tools...
Preparing your chemistry tools...
Draw or paste a fragment — find every compound that contains it. Powered by RDKit, matched on real molecular connectivity, not text.
Find compounds that contain this fragment. Accepts SMILES or SMARTS.
Try an example
Your fragment becomes a query graph. RDKit checks whether it embeds inside each compound’s real bond network — so c1ccccc1 finds every benzene ring, not every "C6H6".
Each molecule is hashed into a Morgan (ECFP) fingerprint. Tanimoto similarity ranks the library by how many structural features it shares with your query.
Every structure’s SMILES is cross-checked against its NIST molecular formula by the calculation engine before it can be searched — no silent bad data.
Common SMILES / SMARTS queries you can paste into the search box
c1ccccc1Aromatic (benzene) ring[OX2H]Hydroxyl group (alcohols, phenols, acids)[CX3](=O)[OX2H1]Carboxylic acid[CX3]=[OX1]Carbonyl (aldehydes, ketones, acids)[NX3;H2]Primary amine[CX3H1]=OAldehydeC(=O)OEster / carboxyl groupC#NNitrileSubstructure search returns compounds that strictly contain your fragment as a subgraph. Similarity search returns compounds that are structurally close overall, ranked by Tanimoto similarity of their fingerprints — useful when there is no exact fragment in common.
A formula like C6H6 tells you the atom counts, not how the atoms are bonded. Substructure matching needs connectivity, which is why this tool searches a curated library of verified SMILES structures rather than the full formula database.
No. You can draw a structure or paste a plain SMILES (e.g. c1ccccc1 for benzene). SMARTS just gives you finer control over functional-group queries like [OX2H] for any hydroxyl group.
Every SMILES in the searchable library is verified at build time: RDKit computes its element composition and it must match the compound’s independently-sourced molecular formula, or the test fails.