Molecules
Search structured molecule pages, starting with reviewed UniProtKB/Swiss-Prot proteins by accession, name, organism or function. Each protein page pairs sequence length, synonyms and biological function with PDB experimental structures, AlphaFold DB predictions, encoding genes and targeting drugs.
Receptor tyrosine kinase binding ligands of the EGF family, with resolved X-ray and cryo-EM structures and linked targeting drugs…
What's in the dataset
UniProt identity and biology
Reviewed UniProtKB/Swiss-Prot accession, protein name, synonyms, organism, canonical sequence length, biological function and ChEMBL target identifiers.
Experimental PDB structures
Resolved 3D structures from the Protein Data Bank with experimental method (X-ray diffraction, cryo-EM, NMR) and resolution in angstroms.
AlphaFold DB predictions
Predicted 3D structures from AlphaFold DB with predictor metadata, mean pLDDT confidence score and direct links to the repository entry.
Encoding genes
Links to each encoding gene page, connecting protein structure and function to target tractability and disease associations.
Targeting drugs
Approved and investigational drugs that act on the protein through its encoding genes, with modality and development stage.
Provenance
Source attribution and retrieval dates across UniProtKB/Swiss-Prot, RCSB PDB and AlphaFold DB.
Search by what matters
Describe a protein or molecule by function, organism or structure method, or search by UniProt accession, ChEMBL target ID or PDB identifier. To keep results to molecule pages, add includeDomains: ["exa.ai/library/molecule"].
Read the docs →# Same endpoint and API key as web search
curl https://api.exa.ai/search \
-H "x-api-key: $EXA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"query": "human kinase proteins with resolved cryo-EM structures",
"includeDomains": ["exa.ai/library/molecule"],
"contents": { "text": true }
}'What teams build with it
Structural biology and drug design
Find proteins with high-resolution PDB structures or high-confidence AlphaFold models alongside their targeting drugs.
Target biology research
Move from a protein's function and 3D structures directly to its encoding gene and clinical drug pipeline.
Grounded answers on molecules
Back answers about protein function, sequence length and target pharmacology with reviewed Swiss-Prot records.
Searchable alongside the general web index
Molecules live in the same index as the open web. There is no separate API or integration. Use the same /search endpoint and API key you already use.
Query molecules on their own, or mix them with papers, genes and drugs in a single call.
Common questions
What's on a molecule page?
Protein records include UniProt accession, name, synonyms, organism, canonical sequence length, biological function, UniProt and ChEMBL identifiers, experimental PDB and predicted AlphaFold DB structures, encoding genes and targeting drugs.
Where does the molecule and structure data come from?
From UniProtKB/Swiss-Prot reviewed entries, Protein Data Bank (PDB) experimental structure records and AlphaFold DB structure predictions.
Are both experimental and predicted structures included?
Yes. Each protein page lists both experimental structures (with method and resolution in angstroms) and AlphaFold DB predicted models (with mean pLDDT confidence scores).
How do molecules link to genes and drugs?
Each protein links to its encoding gene pages in Exa Library and lists the drugs targeting those genes.
Is there a molecule data API?
Yes. Molecules are searchable through the same Exa /search endpoint and API key you use for web search, so there is no separate API to set up.
How do I search only molecules?
Pass includeDomains: ["exa.ai/library/molecule"] to the /search endpoint.
Have data that belongs here?
If you have data that adds to this collection, partner with us. Your data reaches 900K+ developers running 2B+ searches a month.
Start searching molecules
Get an API key in minutes, or talk to us about volume, coverage and enterprise terms.