ChEMBL
LiveChEMBL Bioactivity Database (EMBL-EBI)
ChEMBL is EMBL-EBI's manually-curated database of bioactive drug-like molecules — the compounds, assays, targets, mechanisms and indications behind modern drug discovery, extracted from the medicinal-chemistry literature and other sources. This MCP lets an AI assistant search, retrieve, and aggregate molecules — by protein target, mechanism of action, therapeutic indication, ATC class, development phase, physicochemical properties (molecular weight, Lipinski violations, …), or exact structure (InChIKey / SMILES). Beyond keyword and structured filters, it also searches by meaning — finding molecules by concept (what they do, what they treat, how they act) even when the wording differs.
Connect
https://mcp.polarisray.com/chembl
Transport: http · Auth:
none · See the
connection guide for client-specific steps.
Tools
-
search_compoundsSearch molecules by free text and filters — protein target, mechanism/action type, therapeutic indication, ATC class, molecule type, development phase (approved drugs or a minimum phase), structure (InChIKey/SMILES), physicochemical ranges (molecular weight, Lipinski Ro5 violations), and flags like withdrawn or black-box warning. -
semantic_search_compoundsSearch molecules by meaning, not just exact keywords — finds relevant compounds by concept across names, mechanisms and indications, blended with keyword search. -
get_compoundRetrieve the full record for a single molecule by its ChEMBL ID — structure (SMILES/InChI), computed properties, synonyms, ATC codes, indications, mechanisms, and a per-target activity summary. -
aggregate_compoundsCount and group molecules by target, mechanism, action type, ATC class, molecule type, development phase, indication, or organism — e.g. the most-drugged targets among approved drugs. -
resolveNormalize lay terms, synonyms, and abbreviations to standard concepts (e.g. "COX-2", "blood thinner") for precise filtering.
Example questions
- "Which approved drugs inhibit cyclooxygenase, and what are their mechanisms?"
- "Find small molecules in phase 2+ that target the EGFR protein."
- "What compounds are indicated for Parkinson's disease by mechanism of action?"
- "Look up the molecule with this InChIKey and show its targets and properties."
- "Which human protein targets have the most distinct bioactive compounds?"
About this dataset
ChEMBL is the European Bioinformatics Institute’s manually-curated database of
bioactive molecules with drug-like properties. It brings together chemical
structures, computed physicochemical properties, bioactivity measurements, the
protein (and other) targets those molecules act on, their mechanisms of
action, and their therapeutic indications — the connective tissue between
chemistry and biology that underpins drug discovery. This connector indexes one
record per molecule (keyed by its stable ChEMBL ID), rolling up each molecule’s
targets, mechanisms, indications and activity summary, and is rebuilt from each
quarterly ChEMBL release.
Where DailyMed and openFDA SPL describe approved drug labels, ChEMBL spans the far larger universe of bioactive compounds — most of which are research molecules, not marketed drugs — and connects them to their molecular targets and mechanisms.
What you can ask
Ask plain-English questions and the assistant translates them into structured
searches: filter by protein target (preferred name or ChEMBL ID), mechanism of
action and action type (inhibitor, agonist, …), therapeutic indication, ATC
class, molecule type (small molecule, protein, antibody, …), development phase
(restrict to approved drugs, or a minimum clinical phase), or exact structure by
InChIKey or SMILES; constrain physicochemical properties such as molecular weight
or Lipinski rule-of-five violations; retrieve a full molecule record by its ChEMBL
ID; or aggregate across the corpus to see, for example, which targets have the most
distinct bioactive compounds. Hits include an official url to the ChEMBL
compound page.
Search by meaning
In addition to keyword and structured search, you can search molecules by meaning rather than exact words — over the annotated molecules (those with a name, mechanism, or indication). Ask for compounds that “block inflammation via the COX pathway” or are “used for high blood pressure” and it surfaces conceptually related molecules even when the phrasing differs, with the same structured filters (target, phase, ATC, …) still applying.
Caveats
ChEMBL aggregates bioactivity data from many assays and from the primary literature; values vary in methodology and completeness, and counts should be treated as signals rather than definitive statistics. The vast majority of molecules are research compounds, not approved drugs. ChEMBL data are licensed CC BY-SA 3.0, so attribution to ChEMBL and the release version is required when reusing results. Always confirm details against the primary ChEMBL record, and never treat bioactivity data as medical advice.