MOLECULAR ASSETS
Technology

One scientific interface. Multiple sources of evidence.

This simplified diagram illustrates the major layers of ChemQuery.

From a question to evidence-linked scientific resultsScientific questionNatural language promptStandalone applicationMCP-connected agentsChemQuery / AI layerInterpret · query · analyzeResearch resultsTables · links · exportsCurated scientific data environmentCore curated datasetsPatents · PubChem · ChEMBLClinical trials · FDA · EPANormalized identifiers and joinsSupplemental sourcesAdditional scientific tablesand specialized contentSelected for each questionCustomer-controlled dataPrivate tables and datasetsSubject to agreed access controlsIntegration scope defined jointlyConceptual architecture — implementation and permissions depend on customer environment
Information sources

The data behind ChemQuery.

CURATED PATENTS
>40 billion

associations

138 countries · 58 languages

PubChem
ChEMBL
Clinical trials
FDA datasets
EPA datasets
BindingDB
DrugCentral
User-created datasets
Open Targets Project
Numerous other sources

Many datasets contain multiple tables (ChEMBL alone has more than 80). Roughly 70 tables are used for training, more than 200 supplemental tables are available, and in-house or custom tables can be added.