Source Licensing and Terms

BioMCP itself is MIT-licensed, but the data providers it queries are not all licensed the same way.

Three distinctions matter:

• BioMCP does not vendor, mirror, or ship upstream datasets in the repository.
• BioMCP performs on-demand read-only queries against upstream services.
• Returned records, downloaded full text, saved output, and downstream reuse can still be governed by upstream provider terms.

Use this page for provider terms, licensing, redistribution, and account requirements. Use API Keys for setup steps and Data Sources for runtime behavior, endpoints, rate limits, and operational caveats.

The canonical machine-readable inventory for this page lives in sources.json.

How to read this page

• tier Tier 1 means baseline BioMCP usage works without credentials. Tier 2 means the BioMCP feature needs provider credentials, registration, or a provider-controlled account path. Tier 3 means BioMCP can query the provider, but notable reuse, redistribution, attribution, or policy caveats need explicit attention.
• integration_mode direct_api means BioMCP calls the provider directly. indirect_only means BioMCP only surfaces that provider as provenance through another API payload.
• bioMcp_auth none means BioMCP does not need a key for that provider. optional_env means an environment variable improves quota or access quality but is not required for baseline use. required_env means the feature requires a configured environment variable. not_applicable means there is no standalone BioMCP auth path because the provider is indirect-only.

Summary table

Source	Tier	Mode	BioMCP auth	License / terms summary	Redistribution summary	Terms URL
AlphaGenome	2	direct_api	required_env	custom provider terms; access is gated by Google/DeepMind service controls	do not assume open redistribution rights for returned prediction outputs	https://deepmind.google/science/alphagenome/
cBioPortal	3	direct_api	none	public API with study-specific downstream terms	reuse depends on the specific study or consortium behind each dataset	https://www.cbioportal.org/
CDC CVX/MVX	1	direct_api	none	most CDC website materials are public domain, but attribution, disclaimer, and exceptions for third-party or non-federal content still apply	reuse is generally allowed with CDC attribution and non-endorsement language; avoid CDC logos and review exceptions before republishing	https://www.cdc.gov/other/agencymaterials.html
CDC WONDER VAERS	1	direct_api	none	CDC WONDER data use restrictions require statistical reporting/analysis use and prohibit re-identification attempts	reuse is allowed for statistical reporting and analysis with source attribution, but do not try to identify individuals or publish identifying linkages	https://wonder.cdc.gov/datause.html
ChEMBL	1	direct_api	none	EMBL-EBI open data service; ChEMBL is published for broad reuse	reuse is generally allowed under the provider's open-data terms with attribution where required	https://www.ebi.ac.uk/chembl/
CIViC	1	direct_api	none	open community knowledgebase; CIViC content is published for unrestricted reuse	reuse is broadly permitted; attribution remains best practice	https://civicdb.org/home
ClinGen	1	direct_api	none	public ClinGen curation resources with publication and attribution expectations	generally queryable and reusable, but users should preserve attribution and source context	https://clinicalgenome.org/
ClinicalTrials.gov	1	direct_api	none	U.S. government public information service	records are broadly reusable; preserve identifiers and avoid implying NLM endorsement	https://clinicaltrials.gov/data-api/about-api
ComplexPortal	1	direct_api	none	EMBL-EBI open data service	reuse follows EMBL-EBI resource terms and any embedded third-party source obligations	https://www.ebi.ac.uk/complexportal/
CPIC	1	direct_api	none	CPIC content is published under CC0 with trademark and attribution guidance	content reuse is broadly allowed, but the CPIC mark/logo has separate restrictions	https://cpicpgx.org/license/
DGIdb	1	direct_api	none	open interaction service; aggregated claims may still reflect upstream source terms	treat DGIdb as an aggregation layer and preserve source attribution for underlying claim providers	https://www.dgidb.org/about
DDInter	3	direct_api	none	CC BY-NC-SA 4.0 with an explicit completeness disclaimer; absence from the database does not prove no interaction exists	reuse requires attribution, non-commercial use, and ShareAlike treatment; do not turn missing rows into safety claims	https://ddinter.scbdd.com/terms/
DisGeNET	2	direct_api	required_env	custom provider terms for API and downloads	do not assume unrestricted redistribution; use according to the provider account terms	https://www.disgenet.com/
EMA	1	direct_api	none	EMA website material may be reused with source attribution; third-party content can carry separate rights	EMA-published website data is generally reusable with attribution, but embedded third-party materials may need separate permission	https://www.ema.europa.eu/en/about-us/about-website/legal-notice
Enrichr	1	direct_api	none	open web/API service with citation expectations for Enrichr and its libraries	reuse of results should preserve attribution to Enrichr and the underlying enrichment libraries	https://maayanlab.cloud/Enrichr/
Europe PMC	1	direct_api	none	open literature metadata service; article and full-text licenses vary by record	metadata is broadly reusable, but full text and PDFs remain governed by article-level licenses	https://europepmc.org/RestfulWebService
Figshare	1	direct_api	none	public repository API; each article/file carries its own license metadata	reuse downloaded article assets according to the Figshare item license and preserve Figshare/article provenance	https://figshare.com/terms
gnomAD	3	direct_api	none	Broad Institute data policies with attribution and service-specific conditions	querying is open, but users should review the gnomAD policies before bulk reuse or republishing	https://gnomad.broadinstitute.org/policies
g:Profiler	1	direct_api	none	open enrichment service with provider citation expectations	results are queryable and reusable, but cite g:Profiler and any underlying databases you depend on	https://biit.cs.ut.ee/gprofiler/help.cgi
GTEx	1	direct_api	none	NIH-hosted public-access expression resource	public summary/expression views are broadly reusable; controlled-access data remains outside BioMCP's scope	https://gtexportal.org/home/documentationPage
NCBI Genetic Testing Registry	1	direct_api	none	NLM/NCBI government data service with public bulk-export access	GTR records are broadly queryable and reusable, but preserve accession provenance and review any embedded third-party identifiers separately	https://www.ncbi.nlm.nih.gov/gtr/docs/faq/
GWAS Catalog	1	direct_api	none	EMBL-EBI resource terms; summary statistics may carry separate licenses	query results are generally reusable, but dataset-level summary statistics can have separate downstream terms	https://www.ebi.ac.uk/gwas/docs/about
Human Protein Atlas	3	direct_api	none	CC BY-SA 4.0 for copyrightable parts of the database	reuse is allowed with attribution and ShareAlike; third-party components may impose extra conditions	https://www.proteinatlas.org/about/licence
HPO JAX API	1	direct_api	none	open HPO data with attribution and integrity requirements	reuse is allowed, but users should preserve attribution, version context, and source integrity	https://human-phenotype-ontology.github.io/license.html
InterPro	1	direct_api	none	EMBL-EBI open data resource	reuse follows InterPro/EMBL-EBI resource terms and any embedded member-database obligations	https://www.ebi.ac.uk/interpro/
KEGG	3	direct_api	none	custom KEGG terms; academic users may freely use the website, non-academic use requires a commercial license	do not assume commercial redistribution rights; query access does not grant a redistribution license	https://www.kegg.jp/kegg/legal.html
LitSense2	1	direct_api	none	NCBI/NLM public-domain literature service	query results are broadly reusable, but preserve article-level provenance and record rights separately	https://www.ncbi.nlm.nih.gov/research/litsense2-api/
MedlinePlus	1	direct_api	none	NLM public-information service with trademark and endorsement guidance	content is widely reusable, but preserve attribution and avoid implying MedlinePlus/NLM endorsement	https://medlineplus.gov/about/using/
Monarch Initiative	1	direct_api	none	open integrated knowledge graph; underlying source licenses still matter	results can be queried openly, but downstream reuse should respect the original sources folded into Monarch	https://monarchinitiative.org/
Mutalyzer	1	direct_api	none	public Mutalyzer normalization service; provider terms apply to returned validation output	preserve Mutalyzer source identity and returned HGVS/protein fields when reusing results	https://mutalyzer.nl/
MyChem.info	1	direct_api	none	BioThings aggregation service; upstream source terms continue to apply	do not assume aggregator responses are relicensed; preserve source provenance for downstream reuse	https://docs.mychem.info/en/latest/
MyDisease.info	1	direct_api	none	BioThings aggregation service; source ontologies and datasets keep their own terms	treat payloads as aggregated source data rather than a new umbrella license	https://docs.mydisease.info/en/latest/
MyGene.info	1	direct_api	none	BioThings aggregation service; source-specific terms remain attached to underlying records	reuse should preserve provenance back to NCBI Gene, UniProt, and other upstream sources	https://docs.mygene.info/en/latest/
MyVariant.info	1	direct_api	none	BioThings aggregation service; indirect providers retain their own terms	ClinVar, COSMIC, Cancer Genome Interpreter, and gnomAD-related fields should be treated according to their original providers' terms	https://docs.myvariant.info/en/latest/
NCBI E-utilities	1	direct_api	optional_env	NLM public-domain utility service	utility responses are broadly reusable; article-level full text still follows the returned record's license context	https://www.ncbi.nlm.nih.gov/books/NBK25501/
NCBI ID Converter	1	direct_api	optional_env	NLM public-domain utility service	utility results are broadly reusable; keep article-level identifiers and downstream article licenses distinct	https://pmc.ncbi.nlm.nih.gov/tools/idconv/
NCI CTS	2	direct_api	required_env	custom provider API terms for the NCI Clinical Trials Search API	query output is usable for search and review, but downstream reuse should follow NCI API terms and record provenance	https://clinicaltrialsapi.cancer.gov/
NIH Reporter	1	direct_api	none	NIH-operated public funding-reporting service	project and grant metadata are broadly reusable; preserve NIH Reporter attribution, fiscal-year context, and grant identifiers	https://www.nih.gov/web-policies-notices
OLS4	1	direct_api	none	EMBL-EBI ontology browser; each ontology keeps its own license	ontology metadata is queryable, but downstream reuse depends on the specific ontology surfaced	https://www.ebi.ac.uk/ols4/
OncoKB	2	direct_api	required_env	custom provider terms; academic research access is no-fee but licensed, commercial/clinical use requires a paid license	do not assume open redistribution rights for OncoKB data or proprietary treatment descriptions	https://faq.oncokb.org/licensing
OpenFDA	1	direct_api	optional_env	FDA-origin public data and API terms	data is broadly reusable, but avoid implying FDA endorsement and preserve source context	https://open.fda.gov/apis/authentication/
OpenTargets	1	direct_api	none	Open Targets data is CC0; platform code is Apache 2.0	platform data is dedicated to the public domain, but linked evidence still carries source provenance	https://platform-docs.opentargets.org/licence
PharmGKB	3	direct_api	none	ClinPGx API data is CC BY-SA 4.0 and subject to the provider's data usage policy	reuse is allowed with attribution and ShareAlike; some underlying annotations and external assets may add extra constraints	https://api.pharmgkb.org/
PMC OA	1	direct_api	optional_env	open-access subset only; article licenses vary within PMC OA	full text is reusable only according to each article's specific PMC Open Access license	https://pmc.ncbi.nlm.nih.gov/tools/openftlist/
PubMed	1	direct_api	optional_env	NLM public-domain search and metadata service	search results are broadly reusable, but article-level abstracts, full text, and downstream reuse still depend on the returned record context	https://www.ncbi.nlm.nih.gov/books/NBK25501/
PubTator3	1	direct_api	optional_env	NCBI/NLM public-domain annotation service	results are broadly reusable, but preserve PMID/source provenance and article-level rights separately	https://www.ncbi.nlm.nih.gov/research/pubtator3/api
QuickGO	1	direct_api	none	GO/EMBL-EBI open data service	query results are generally reusable; preserve GO/EMBL-EBI attribution where expected	https://www.ebi.ac.uk/QuickGO/
Reactome	1	direct_api	none	Reactome pathway content is CC BY 4.0, with some data exports additionally placed under CC0	reuse is allowed with attribution; preserve pathway/source provenance in downstream materials	https://reactome.org/license
SEER Explorer	1	direct_api	none	U.S. government cancer-statistics website and public explorer	summary survival statistics are broadly reusable, but users should preserve SEER/NCI attribution and understand the explorer endpoints are undocumented UI routes	https://seer.cancer.gov/about/using-website.html
Semantic Scholar	2	direct_api	optional_env	custom API license agreement	the API license restricts repackaging, resale, and broad commercial redistribution without expanded licensing	https://www.semanticscholar.org/product/api/license
STRING	1	direct_api	none	CC BY 4.0	reuse is allowed with attribution to STRING and the original publication/resource	https://string-db.org/cgi/access?footer_active_subpage=licensing
UMLS	2	direct_api	required_env	custom UMLS Metathesaurus license and terminology-specific appendices	do not assume unrestricted redistribution; some embedded vocabularies add their own restrictions or affiliate licenses	https://www.nlm.nih.gov/databases/umls.html
UniProt	1	direct_api	none	CC BY 4.0	reuse is allowed with attribution; linked cross-references can have their own terms	https://www.uniprot.org/help/license
VariantValidator	1	direct_api	none	public VariantValidator service; provider terms apply to returned validation output	preserve VariantValidator source identity, warnings, and genomic descriptions when reusing results	https://variantvalidator.org/
WHO Prequalification	1	direct_api	none	WHO content generally requires attribution; commercial reuse may require permission and third-party materials can carry separate rights	preserve WHO attribution, review commercial-use conditions, and check third-party material rights before republishing	https://www.who.int/about/policies/publishing/copyright
WikiPathways	1	direct_api	none	CC0	pathway content is dedicated to the public domain; attribution is still good scholarly practice	https://classic.wikipathways.org/index.php/WikiPathways:License_Terms
AlphaFold DB	1	indirect_only	not_applicable	AlphaFold DB structural predictions are published for broad open use	reuse is generally open, but preserve model/source provenance and article citations	https://alphafold.ebi.ac.uk/faq
Cancer Genome Interpreter	3	indirect_only	not_applicable	custom tool terms	do not assume commercial reuse rights; the official terms restrict some external and commercial use	https://www.cancergenomeinterpreter.org/conditions
ClinVar	1	indirect_only	not_applicable	NCBI public-domain submission archive	records are broadly reusable, but preserve accession/provenance and submitter context	https://www.ncbi.nlm.nih.gov/clinvar/docs/maintenance_use/
COSMIC	3	indirect_only	not_applicable	custom COSMIC licensing with commercial restrictions	direct redistribution and direct integration remain intentionally unsupported without a separate COSMIC license	https://www.sanger.ac.uk/legal/cosmic-licensing/
Disease Ontology	1	indirect_only	not_applicable	open disease ontology project	reuse is generally open; preserve ontology version and source references	https://disease-ontology.org/
DrugBank	3	indirect_only	not_applicable	custom DrugBank terms of use and licensing	use or redistribution of DrugBank content requires a DrugBank license; do not assume open downstream rights	https://trust.drugbank.com/drugbank-trust-center/drugbank-terms-of-service
Drugs@FDA	1	indirect_only	not_applicable	FDA-origin public information	approval records are broadly reusable; avoid implying FDA endorsement	https://open.fda.gov/apis/drug/drugsfda/
MONDO	1	indirect_only	not_applicable	CC BY 4.0	reuse is allowed with attribution and ontology version tracking	https://mondo.monarchinitiative.org/pages/download/
PDB	1	indirect_only	not_applicable	PDB archive data is CC0 1.0	data is broadly reusable; attribution to original structure authors is encouraged	https://www.rcsb.org/pages/usage-policy

Tier 1 - Baseline use without credentials

ChEMBL

• BioMCP surfaces: get drug <name> targets; get drug <name> indications
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: EMBL-EBI open data service; ChEMBL is published for broad reuse
• Redistribution / reuse summary: reuse is generally allowed under the provider's open-data terms with attribution where required
• Official terms URL: https://www.ebi.ac.uk/chembl/
• Reviewed on: 2026-03-20
• Notes: get drug <name> targets keeps ChEMBL as the generic target/mechanism source; CIViC variant labels are rendered separately when present.

CIViC

• BioMCP surfaces: get variant <id> civic; get gene <symbol> civic; get drug <name> targets; get drug <name> civic; get disease <id> civic; get disease <id> variants
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: open community knowledgebase; CIViC content is published for unrestricted reuse
• Redistribution / reuse summary: reuse is broadly permitted; attribution remains best practice
• Official terms URL: https://civicdb.org/home
• Reviewed on: 2026-03-20
• Notes: CIViC is treated here as an open-access evidence source surfaced directly by BioMCP. Drug target output may add a separate CIViC variant-target annotation line without merging those labels into the generic targets section.

ClinGen

• BioMCP surfaces: get gene <symbol> clingen
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: public web/API access
• License / terms summary: public ClinGen curation resources with publication and attribution expectations
• Redistribution / reuse summary: generally queryable and reusable, but users should preserve attribution and source context
• Official terms URL: https://clinicalgenome.org/
• Reviewed on: 2026-03-20
• Notes: Open Targets currently lists ClinGen under CC0 for its own ingestion, but BioMCP links to ClinGen's official project site because that is the provider surface users encounter directly.

ClinicalTrials.gov

• BioMCP surfaces: search trial; get trial <nct_id>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: U.S. government public information service
• Redistribution / reuse summary: records are broadly reusable; preserve identifiers and avoid implying NLM endorsement
• Official terms URL: https://clinicaltrials.gov/data-api/about-api
• Reviewed on: 2026-03-20
• Notes: BioMCP uses the public v2 API as the baseline trial backend.

ComplexPortal

• BioMCP surfaces: get protein <id> complexes
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: EMBL-EBI open data service
• Redistribution / reuse summary: reuse follows EMBL-EBI resource terms and any embedded third-party source obligations
• Official terms URL: https://www.ebi.ac.uk/complexportal/
• Reviewed on: 2026-03-20
• Notes: BioMCP calls the IntAct Complex Portal web service. Complex membership data is queried on demand only.

CPIC

• BioMCP surfaces: search pgx; get pgx <gene_or_drug>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: CPIC content is published under CC0 with trademark and attribution guidance
• Redistribution / reuse summary: content reuse is broadly allowed, but the CPIC mark/logo has separate restrictions
• Official terms URL: https://cpicpgx.org/license/
• Reviewed on: 2026-03-20
• Notes: CPIC announced in March 2026 that content is moving to ClinPGx, but current CPIC URLs continue to resolve.

DGIdb

• BioMCP surfaces: get gene <symbol> interactions; get drug <name> interactions
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: open interaction service; aggregated claims may still reflect upstream source terms
• Redistribution / reuse summary: treat DGIdb as an aggregation layer and preserve source attribution for underlying claim providers
• Official terms URL: https://www.dgidb.org/about
• Reviewed on: 2026-03-20
• Notes: DGIdb itself is open to query, but it aggregates claims from many external drug-gene sources.

CDC CVX/MVX

• BioMCP surfaces: search drug <name> --region eu|all; search drug <name> --region who --product-type vaccine; search drug <name> (default plain-name all-region search); biomcp health; biomcp cvx sync
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: on-demand local download by BioMCP on first EU/default/WHO-vaccine bridge use, or manual preseed via BIOMCP_CVX_DIR
• License / terms summary: most CDC website materials are public domain, but attribution, disclaimer, and exceptions for third-party or non-federal content still apply
• Redistribution / reuse summary: reuse is generally allowed with CDC attribution and non-endorsement language; avoid CDC logos and review exceptions before republishing
• Official terms URL: https://www.cdc.gov/other/agencymaterials.html
• Reviewed on: 2026-04-17
• Notes: BioMCP auto-downloads cvx.txt, TRADENAME.txt, and mvx.txt into BIOMCP_CVX_DIR or the default data directory on first use, refreshes stale files after 30 days, and supports explicit refresh via biomcp cvx sync. The bundle augments EMA/default vaccine search plus explicit WHO vaccine name/brand search after MyChem misses; pure --region us search and WHO finished-pharma/API lookups do not use it.

CDC WONDER VAERS

• BioMCP surfaces: search adverse-event --source vaers; search adverse-event --source all; biomcp health
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public CDC WONDER web/API access; consent to the CDC WONDER data use restrictions is required
• License / terms summary: CDC WONDER data use restrictions require statistical reporting/analysis use and prohibit re-identification attempts
• Redistribution / reuse summary: reuse is allowed for statistical reporting and analysis with source attribution, but do not try to identify individuals or publish identifying linkages
• Official terms URL: https://wonder.cdc.gov/datause.html
• Reviewed on: 2026-04-18
• Notes: BioMCP calls the CDC WONDER VAERS D8 XML POST endpoint for aggregate vaccine adverse-event summaries only. The source is aggregate-only in this ticket; case-level VAERS export and direct report detail retrieval are intentionally out of scope.

EMA

• BioMCP surfaces: search drug --region eu|all; get drug <name> regulatory|safety|shortage --region eu|all
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: on-demand local download by BioMCP on first EMA use, or manual preseed via BIOMCP_EMA_DIR
• License / terms summary: EMA website material may be reused with source attribution; third-party content can carry separate rights
• Redistribution / reuse summary: EMA-published website data is generally reusable with attribution, but embedded third-party materials may need separate permission
• Official terms URL: https://www.ema.europa.eu/en/about-us/about-website/legal-notice
• Reviewed on: 2026-03-24
• Notes: BioMCP auto-downloads the EMA human-medicines JSON batch into BIOMCP_EMA_DIR or the default data directory on first use, refreshes stale files, and supports explicit refresh via biomcp ema sync. The download landing page is https://www.ema.europa.eu/en/about-us/about-website/download-website-data-json-data-format.

WHO Prequalification

• BioMCP surfaces: search drug --region who|all; search drug --indication <disease> --region who; search drug <name> --region who --product-type <finished_pharma|api|vaccine>; get drug <name> regulatory --region who|all
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: on-demand local download by BioMCP on first WHO use, or manual preseed via BIOMCP_WHO_DIR
• License / terms summary: WHO content generally requires attribution; commercial reuse may require permission and third-party materials can carry separate rights
• Redistribution / reuse summary: preserve WHO attribution, review commercial-use conditions, and check third-party material rights before republishing
• Official terms URL: https://www.who.int/about/policies/publishing/copyright
• Reviewed on: 2026-04-10
• Notes: BioMCP auto-downloads the WHO finished-pharmaceutical-products CSV, active-pharmaceutical-ingredients CSV, and vaccine CSV into BIOMCP_WHO_DIR or the default data directory on first use, refreshes stale files, supports WHO-only --product-type <finished_pharma|api|vaccine> filtering on explicit --region who searches, and supports explicit refresh via biomcp who sync. WHO vaccine support is search-only; get drug <name> regulatory --region who|all remains finished-pharma/API only.

NCBI Genetic Testing Registry

• BioMCP surfaces: search diagnostic; get diagnostic <gtr_accession> [genes|conditions|methods|regulatory]; biomcp health; biomcp gtr sync
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: on-demand local download by BioMCP on first diagnostic use, or manual preseed via BIOMCP_GTR_DIR
• License / terms summary: NLM/NCBI government data service with public bulk-export access
• Redistribution / reuse summary: GTR records are broadly queryable and reusable, but preserve accession provenance and review any embedded third-party identifiers separately
• Official terms URL: https://www.ncbi.nlm.nih.gov/gtr/docs/faq/
• Reviewed on: 2026-04-17
• Notes: BioMCP auto-downloads test_version.gz and test_condition_gene.txt into BIOMCP_GTR_DIR or the default data directory on first use, refreshes stale files after 7 days, and supports explicit refresh via biomcp gtr sync.

WHO Prequalified IVD

• BioMCP surfaces: search diagnostic --source who-ivd|all; get diagnostic <who_ivd_product_code> [conditions|regulatory]; biomcp health; biomcp who-ivd sync
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: on-demand local download by BioMCP on first WHO IVD use, or manual preseed via BIOMCP_WHO_IVD_DIR
• License / terms summary: WHO content generally requires attribution; commercial reuse may require permission and third-party materials can carry separate rights
• Redistribution / reuse summary: preserve WHO attribution, review commercial-use conditions, and check third-party material rights before republishing WHO IVD-derived content
• Official terms URL: https://www.who.int/about/policies/publishing/copyright
• Reviewed on: 2026-04-18
• Notes: BioMCP auto-downloads who_ivd.csv into BIOMCP_WHO_IVD_DIR or the default data directory on first use, refreshes stale files after 72 hours, and supports explicit refresh via biomcp who-ivd sync.

Enrichr

• BioMCP surfaces: get gene <symbol> ontology
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: open web/API service with citation expectations for Enrichr and its libraries
• Redistribution / reuse summary: reuse of results should preserve attribution to Enrichr and the underlying enrichment libraries
• Official terms URL: https://maayanlab.cloud/Enrichr/
• Reviewed on: 2026-03-20
• Notes: Gene enrichment sections inside BioMCP use Enrichr; top-level biomcp enrich uses g:Profiler instead.

Europe PMC

• BioMCP surfaces: search article; get article <pmid>; get article <id> fulltext
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: open literature metadata service; article and full-text licenses vary by record
• Redistribution / reuse summary: metadata is broadly reusable, but full text and PDFs remain governed by article-level licenses
• Official terms URL: https://europepmc.org/RestfulWebService
• Reviewed on: 2026-03-20
• Notes: BioMCP uses Europe PMC for search, bibliographic metadata, and article full-text XML lookups. Open-access reuse depends on the publication license attached to each record.

Figshare

• BioMCP surfaces: get article <id> assets; get article <id> asset <name>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: public repository API; each article/file carries its own license metadata
• Redistribution / reuse summary: reuse downloaded article assets according to the Figshare item license and preserve Figshare/article provenance
• Official terms URL: https://figshare.com/terms
• Reviewed on: 2026-06-06
• Notes: BioMCP uses Figshare only as an article-asset fallback after PMC OA assets are unavailable and Semantic Scholar discovery points at a supported Figshare/AACR Figshare article URL. BioMCP downloads bytes from the Figshare API download_url and does not parse supplement contents.

g:Profiler

• BioMCP surfaces: enrich <GENE1,GENE2,...>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: open enrichment service with provider citation expectations
• Redistribution / reuse summary: results are queryable and reusable, but cite g:Profiler and any underlying databases you depend on
• Official terms URL: https://biit.cs.ut.ee/gprofiler/help.cgi
• Reviewed on: 2026-03-20
• Notes: BioMCP uses g:Profiler only for top-level gene-set enrichment.

GTEx

• BioMCP surfaces: get gene <symbol> expression
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public portal/API for public-access expression data
• License / terms summary: NIH-hosted public-access expression resource
• Redistribution / reuse summary: public summary/expression views are broadly reusable; controlled-access data remains outside BioMCP's scope
• Official terms URL: https://gtexportal.org/home/documentationPage
• Reviewed on: 2026-03-20
• Notes: BioMCP only queries public GTEx expression endpoints, not controlled-access donor-level data.

GWAS Catalog

• BioMCP surfaces: search gwas; get variant <id> gwas
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: EMBL-EBI resource terms; summary statistics may carry separate licenses
• Redistribution / reuse summary: query results are generally reusable, but dataset-level summary statistics can have separate downstream terms
• Official terms URL: https://www.ebi.ac.uk/gwas/docs/about
• Reviewed on: 2026-03-20
• Notes: Open Targets reports GWAS Catalog summary statistics under CC0 while the service remains under EMBL-EBI terms.

HPO JAX API

• BioMCP surfaces: search phenotype; discover phenotype term resolution
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public ontology API
• License / terms summary: open HPO data with attribution and integrity requirements
• Redistribution / reuse summary: reuse is allowed, but users should preserve attribution, version context, and source integrity
• Official terms URL: https://human-phenotype-ontology.github.io/license.html
• Reviewed on: 2026-03-20
• Notes: The HPO project's own license page is more specific than generic JAX site text and is the clearest official usage statement currently exposed.

InterPro

• BioMCP surfaces: get protein <id> domains
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: EMBL-EBI open data resource
• Redistribution / reuse summary: reuse follows InterPro/EMBL-EBI resource terms and any embedded member-database obligations
• Official terms URL: https://www.ebi.ac.uk/interpro/
• Reviewed on: 2026-03-20
• Notes: InterPro aggregates signatures from multiple member databases; downstream interpretation should keep that provenance.

MedlinePlus

• BioMCP surfaces: discover plain-language topics; get disease <name_or_id> clinical_features
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public search API
• License / terms summary: NLM public-information service with trademark and endorsement guidance
• Redistribution / reuse summary: content is widely reusable, but preserve attribution and avoid implying MedlinePlus/NLM endorsement
• Official terms URL: https://medlineplus.gov/about/using/
• Reviewed on: 2026-03-20
• Notes: BioMCP uses MedlinePlus for best-effort plain-language discover context and opt-in disease clinical-feature summaries for configured diseases.

Monarch Initiative

• BioMCP surfaces: get disease <id> genes; get disease <id> models; search phenotype
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: open integrated knowledge graph; underlying source licenses still matter
• Redistribution / reuse summary: results can be queried openly, but downstream reuse should respect the original sources folded into Monarch
• Official terms URL: https://monarchinitiative.org/
• Reviewed on: 2026-03-20
• Notes: BioMCP uses Monarch as an aggregator for disease, phenotype, and model-organism relationships rather than as the legal source of every embedded assertion.

Mutalyzer

• BioMCP surfaces: variant normalize
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: public Mutalyzer normalization service; provider terms apply to returned validation output
• Redistribution / reuse summary: preserve Mutalyzer source identity and returned HGVS/protein fields when reusing results
• Official terms URL: https://mutalyzer.nl/
• Reviewed on: 2026-05-22
• Notes: BioMCP calls GET /normalize/{description} for explicit transcript HGVS and preserves source warnings/status without classifying variants.

MyChem.info

• BioMCP surfaces: search drug; get drug <name>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: BioThings aggregation service; upstream source terms continue to apply
• Redistribution / reuse summary: do not assume aggregator responses are relicensed; preserve source provenance for downstream reuse
• Official terms URL: https://docs.mychem.info/en/latest/
• Reviewed on: 2026-03-20
• Notes: DrugBank and other upstream providers appear in MyChem payloads with their own terms.

MyDisease.info

• BioMCP surfaces: search disease; get disease <id>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: BioThings aggregation service; source ontologies and datasets keep their own terms
• Redistribution / reuse summary: treat payloads as aggregated source data rather than a new umbrella license
• Official terms URL: https://docs.mydisease.info/en/latest/
• Reviewed on: 2026-03-20
• Notes: Disease Ontology and MONDO appear through MyDisease.info as indirect provenance sources.

MyGene.info

• BioMCP surfaces: search gene; get gene <symbol>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: BioThings aggregation service; source-specific terms remain attached to underlying records
• Redistribution / reuse summary: reuse should preserve provenance back to NCBI Gene, UniProt, and other upstream sources
• Official terms URL: https://docs.mygene.info/en/latest/
• Reviewed on: 2026-03-20
• Notes: BioMCP uses MyGene.info mainly as an identity/normalization layer rather than as the legal origin of all gene data.

MyVariant.info

• BioMCP surfaces: search variant; get variant <id>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: BioThings aggregation service; indirect providers retain their own terms
• Redistribution / reuse summary: ClinVar, COSMIC, Cancer Genome Interpreter, and gnomAD-related fields should be treated according to their original providers' terms
• Official terms URL: https://docs.myvariant.info/en/latest/
• Reviewed on: 2026-03-20
• Notes: BioMCP intentionally documents several indirect-only provenance rows that arrive through MyVariant.info payloads.

NCBI ID Converter

• BioMCP surfaces: get article <id> fulltext
• Integration mode: direct_api
• BioMCP auth: optional_env via NCBI_API_KEY
• Provider access / registration: open public utility; optional My NCBI API key improves throughput
• License / terms summary: NLM public-domain utility service
• Redistribution / reuse summary: utility results are broadly reusable; keep article-level identifiers and downstream article licenses distinct
• Official terms URL: https://pmc.ncbi.nlm.nih.gov/tools/idconv/
• API key / account URL: https://www.ncbi.nlm.nih.gov/account/settings/
• Reviewed on: 2026-03-20
• Notes: NCBI API keys increase throughput but are not required for baseline BioMCP usage.

NCBI E-utilities

• BioMCP surfaces: get article <id> fulltext
• Integration mode: direct_api
• BioMCP auth: optional_env via NCBI_API_KEY
• Provider access / registration: open public E-utilities endpoint; optional My NCBI API key improves throughput
• License / terms summary: NLM public-domain utility service
• Redistribution / reuse summary: utility responses are broadly reusable; article-level full text still follows the returned record's license context
• Official terms URL: https://www.ncbi.nlm.nih.gov/books/NBK25501/
• API key / account URL: https://www.ncbi.nlm.nih.gov/account/settings/
• Reviewed on: 2026-04-10
• Notes: BioMCP uses PMC efetch as one XML full-text fallback when Europe PMC does not serve the article; PMC OA and PMC article HTML are separate later fallback surfaces. NCBI_API_KEY raises the baseline NCBI E-utilities budget but is not required.

NIH Reporter

• BioMCP surfaces: get gene <symbol> funding; get disease <name_or_id> funding
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: NIH-operated public funding-reporting service
• Redistribution / reuse summary: project and grant metadata are broadly reusable; preserve NIH Reporter attribution, fiscal-year context, and grant identifiers
• Official terms URL: https://www.nih.gov/web-policies-notices
• Reviewed on: 2026-04-11
• Notes: BioMCP uses the public NIH Reporter v2 projects search API for exact-phrase title/abstract funding lookups. Funding remains opt-in on gene and disease cards rather than being folded into all.

OLS4

• BioMCP surfaces: discover <query>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public ontology service
• License / terms summary: EMBL-EBI ontology browser; each ontology keeps its own license
• Redistribution / reuse summary: ontology metadata is queryable, but downstream reuse depends on the specific ontology surfaced
• Official terms URL: https://www.ebi.ac.uk/ols4/
• Reviewed on: 2026-03-20
• Notes: OLS4 is BioMCP's required backbone for discover. It is an ontology index, not a single-license data source.

OpenFDA

• BioMCP surfaces: search adverse-event; get drug <name> label; get drug <name> approvals; get diagnostic <diagnostic_id> regulatory
• Integration mode: direct_api
• BioMCP auth: optional_env via OPENFDA_API_KEY
• Provider access / registration: open public API; optional key increases quota headroom
• License / terms summary: FDA-origin public data and API terms
• Redistribution / reuse summary: data is broadly reusable, but avoid implying FDA endorsement and preserve source context
• Official terms URL: https://open.fda.gov/apis/authentication/
• API key / account URL: https://open.fda.gov/apis/authentication/
• Reviewed on: 2026-04-18
• Notes: The authentication page says API keys are required while also publishing no-key quotas. BioMCP documents the real runtime behavior: baseline use works without a key, with higher quotas when one is configured. Diagnostic regulatory uses the OpenFDA device 510(k) and PMA endpoints as an exact-name-first live overlay rather than a full mirrored device catalog.

OpenTargets

• BioMCP surfaces: get gene <symbol> diseases; get drug <name> targets; get disease <id> genes
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public GraphQL API
• License / terms summary: Open Targets data is CC0; platform code is Apache 2.0
• Redistribution / reuse summary: platform data is dedicated to the public domain, but linked evidence still carries source provenance
• Official terms URL: https://platform-docs.opentargets.org/licence
• Reviewed on: 2026-03-20
• Notes: get drug <name> targets keeps Open Targets in the generic target section while CIViC variant annotations, when present, are labeled separately.

PMC OA

• BioMCP surfaces: get article <id> fulltext; get article <id> assets; get article <id> asset <name>
• Integration mode: direct_api
• BioMCP auth: optional_env via NCBI_API_KEY
• Provider access / registration: open public utility for the PMC Open Access subset
• License / terms summary: open-access subset only; article licenses vary within PMC OA
• Redistribution / reuse summary: full text and assets are reusable only according to each article's specific PMC Open Access license
• Official terms URL: https://pmc.ncbi.nlm.nih.gov/tools/openftlist/
• API key / account URL: https://www.ncbi.nlm.nih.gov/account/settings/
• Reviewed on: 2026-03-20
• Notes: BioMCP queries PMC OA on demand as one XML full-text rung and as the preferred article-asset provider; it does not ship the article corpus. PMC article HTML is a separate derived fallback, and returned full text/assets are still governed by article-level licenses.

PubMed

• BioMCP surfaces: search article; get article <id>
• Integration mode: direct_api
• BioMCP auth: optional_env via NCBI_API_KEY
• Provider access / registration: open public search/metadata service; optional My NCBI API key improves throughput
• License / terms summary: NLM public-domain search and metadata service
• Redistribution / reuse summary: search results are broadly reusable, but article-level abstracts, full text, and downstream reuse still depend on the returned record context
• Official terms URL: https://www.ncbi.nlm.nih.gov/books/NBK25501/
• API key / account URL: https://www.ncbi.nlm.nih.gov/account/settings/
• Reviewed on: 2026-04-10
• Notes: BioMCP uses PubMed as a first-class article search source while keeping separate rows for PubTator3 annotations, NCBI E-utilities/PMC OA XML full text, NCBI ID Converter bridging, and Semantic Scholar PDF metadata; PMC article HTML is documented as a PMC web fallback in the data-source matrix.

PubTator3

• BioMCP surfaces: search article; get article <pmid> annotations
• Integration mode: direct_api
• BioMCP auth: optional_env via NCBI_API_KEY
• Provider access / registration: open public API; optional My NCBI API key improves throughput
• License / terms summary: NCBI/NLM public-domain annotation service
• Redistribution / reuse summary: results are broadly reusable, but preserve PMID/source provenance and article-level rights separately
• Official terms URL: https://www.ncbi.nlm.nih.gov/research/pubtator3/api
• API key / account URL: https://www.ncbi.nlm.nih.gov/account/settings/
• Reviewed on: 2026-03-20
• Notes: BioMCP uses PubTator3 for article search fan-out and article annotation. NCBI_API_KEY is optional quota uplift only.

QuickGO

• BioMCP surfaces: get gene <symbol> go
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: GO/EMBL-EBI open data service
• Redistribution / reuse summary: query results are generally reusable; preserve GO/EMBL-EBI attribution where expected
• Official terms URL: https://www.ebi.ac.uk/QuickGO/
• Reviewed on: 2026-03-20
• Notes: QuickGO exposes GO data and annotations; some embedded evidence sources can carry their own provenance requirements.

Reactome

• BioMCP surfaces: search pathway; get pathway <id>; get gene <symbol> pathways
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: Reactome pathway content is CC BY 4.0, with some data exports additionally placed under CC0
• Redistribution / reuse summary: reuse is allowed with attribution; preserve pathway/source provenance in downstream materials
• Official terms URL: https://reactome.org/license
• Reviewed on: 2026-03-20
• Notes: Reactome announced in 2017 that some annotation files moved to CC0 while core site/code materials remained under CC BY 4.0.

SEER Explorer

• BioMCP surfaces: get disease <id> survival
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: public website endpoints; no BioMCP account or key path
• License / terms summary: U.S. government cancer-statistics website and public explorer
• Redistribution / reuse summary: summary survival statistics are broadly reusable, but users should preserve SEER/NCI attribution and understand the explorer endpoints are undocumented UI routes
• Official terms URL: https://seer.cancer.gov/about/using-website.html
• Reviewed on: 2026-04-10
• Notes: BioMCP uses the public SEER Explorer site catalog and explorer endpoints for cancer survival summaries. The shipped contract treats malformed or mismatched explorer payloads as temporary unavailability rather than trustworthy data.

STRING

• BioMCP surfaces: get gene <symbol> interactions; get protein <id> interactions
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: CC BY 4.0
• Redistribution / reuse summary: reuse is allowed with attribution to STRING and the original publication/resource
• Official terms URL: https://string-db.org/cgi/access?footer_active_subpage=licensing
• Reviewed on: 2026-03-20
• Notes: BioMCP queries STRING network endpoints on demand and does not package STRING interaction datasets.

UniProt

• BioMCP surfaces: get protein <id>; get gene <symbol> protein
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: CC BY 4.0
• Redistribution / reuse summary: reuse is allowed with attribution; linked cross-references can have their own terms
• Official terms URL: https://www.uniprot.org/help/license
• Reviewed on: 2026-03-20
• Notes: BioMCP also surfaces UniProt cross-references to PDB and AlphaFold DB rather than mirroring those datasets directly.

VariantValidator

• BioMCP surfaces: variant normalize
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: public VariantValidator service; provider terms apply to returned validation output
• Redistribution / reuse summary: preserve VariantValidator source identity, warnings, and genomic descriptions when reusing results
• Official terms URL: https://variantvalidator.org/
• Reviewed on: 2026-05-22
• Notes: BioMCP calls /VariantValidator/variantvalidator/{genome_build}/{variant_description}/{select_transcripts} for explicit transcript HGVS and preserves TranscriptVersionWarning messages without choosing transcripts.

WikiPathways

• BioMCP surfaces: search pathway; get pathway <id>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: CC0
• Redistribution / reuse summary: pathway content is dedicated to the public domain; attribution is still good scholarly practice
• Official terms URL: https://classic.wikipathways.org/index.php/WikiPathways:License_Terms
• Reviewed on: 2026-03-20
• Notes: The current license statement is still hosted on the WikiPathways classic site.

Tier 2 - Credential, account, or license required for the BioMCP feature

AlphaGenome

• BioMCP surfaces: get variant <id> predict
• Integration mode: direct_api
• BioMCP auth: required_env via ALPHAGENOME_API_KEY
• Provider access / registration: provider-controlled access to the AlphaGenome service
• License / terms summary: custom provider terms; access is gated by Google/DeepMind service controls
• Redistribution / reuse summary: do not assume open redistribution rights for returned prediction outputs
• Official terms URL: https://deepmind.google/science/alphagenome/
• API key / account URL: https://deepmind.google/science/alphagenome/
• Reviewed on: 2026-03-20
• Notes: BioMCP calls the hosted gRPC endpoint directly. The public product page is the closest official access reference currently exposed.

DisGeNET

• BioMCP surfaces: get gene <symbol> disgenet; get disease <id> disgenet
• Integration mode: direct_api
• BioMCP auth: required_env via DISGENET_API_KEY
• Provider access / registration: account registration and API key required
• License / terms summary: custom provider terms for API and downloads
• Redistribution / reuse summary: do not assume unrestricted redistribution; use according to the provider account terms
• Official terms URL: https://www.disgenet.com/
• API key / account URL: https://www.disgenet.com/
• Reviewed on: 2026-03-20
• Notes: DisGeNET's public site advertises free and commercial plans. BioMCP documents the API as key-gated and treats it as provider-controlled.

NCI CTS

• BioMCP surfaces: search trial --source nci; get trial <nct_id> --source nci
• Integration mode: direct_api
• BioMCP auth: required_env via NCI_API_KEY
• Provider access / registration: API key required
• License / terms summary: custom provider API terms for the NCI Clinical Trials Search API
• Redistribution / reuse summary: query output is usable for search and review, but downstream reuse should follow NCI API terms and record provenance
• Official terms URL: https://clinicaltrialsapi.cancer.gov/
• API key / account URL: https://clinicaltrialsapi.cancer.gov/
• Reviewed on: 2026-03-20
• Notes: BioMCP treats NCI CTS as an alternate oncology-focused trial backend, not the default public trial source.

OncoKB

• BioMCP surfaces: variant oncokb <id>
• Integration mode: direct_api
• BioMCP auth: required_env via ONCOKB_TOKEN
• Provider access / registration: registration and provider-controlled API access
• License / terms summary: custom provider terms; academic research access is no-fee but licensed, commercial/clinical use requires a paid license
• Redistribution / reuse summary: do not assume open redistribution rights for OncoKB data or proprietary treatment descriptions
• Official terms URL: https://faq.oncokb.org/licensing
• API key / account URL: https://www.oncokb.org/account/register
• Reviewed on: 2026-03-20
• Notes: The FAQ states that programmatic access for academic use still requires registration and provider approval.

Semantic Scholar

• BioMCP surfaces: search article; get article <id> tldr; get article <id> fulltext --pdf; article citations <id>; article references <id>; article recommendations <id>
• Integration mode: direct_api
• BioMCP auth: optional_env via S2_API_KEY
• Provider access / registration: unauthenticated access is available on the shared pool; authenticated access uses the provider-issued API key and remains governed by the API license agreement
• License / terms summary: custom API license agreement
• Redistribution / reuse summary: the API license restricts repackaging, resale, and broad commercial redistribution without expanded licensing
• Official terms URL: https://www.semanticscholar.org/product/api/license
• API key / account URL: https://www.semanticscholar.org/product/api
• Reviewed on: 2026-03-20
• Notes: BioMCP can call Semantic Scholar without S2_API_KEY, but uses a more conservative shared-pool rate limit and recommends the key for dedicated quota and reliability. For get article <id> fulltext --pdf, BioMCP uses Semantic Scholar openAccessPdf metadata, then fetches the third-party PDF URL only after explicit PDF opt-in; the PDF's article-level reuse terms remain separate.

UMLS

• BioMCP surfaces: discover <query> crosswalk enrichment
• Integration mode: direct_api
• BioMCP auth: required_env via UMLS_API_KEY
• Provider access / registration: UTS account plus acceptance of the UMLS Metathesaurus license
• License / terms summary: custom UMLS Metathesaurus license and terminology-specific appendices
• Redistribution / reuse summary: do not assume unrestricted redistribution; some embedded vocabularies add their own restrictions or affiliate licenses
• Official terms URL: https://www.nlm.nih.gov/databases/umls.html
• API key / account URL: https://uts.nlm.nih.gov/uts/signup-login
• Reviewed on: 2026-03-20
• Notes: The UMLS landing page explicitly states that you must accept the license and create a UTS account for access.

Tier 3 - Open or queryable, but with notable terms

cBioPortal

• BioMCP surfaces: get variant <id> cohort; study download; study query
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: public API access for open studies
• License / terms summary: public API with study-specific downstream terms
• Redistribution / reuse summary: reuse depends on the specific study or consortium behind each dataset
• Official terms URL: https://www.cbioportal.org/
• Reviewed on: 2026-03-20
• Notes: BioMCP uses the public API. TCGA-derived studies are broadly open, while some cohorts such as AACR Project GENIE carry additional access or reuse conditions.

DDInter

• BioMCP surfaces: biomcp drug interactions <name>; get drug <name> interactions; biomcp health; biomcp ddinter sync
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: on-demand local download by BioMCP on first interaction use, or manual preseed via BIOMCP_DDINTER_DIR
• License / terms summary: CC BY-NC-SA 4.0 with an explicit completeness disclaimer; absence from the database does not prove no interaction exists
• Redistribution / reuse summary: reuse requires attribution, non-commercial use, and ShareAlike treatment; do not turn missing rows into safety claims or assume commercial redistribution rights
• Official terms URL: https://ddinter.scbdd.com/terms/
• Reviewed on: 2026-04-25
• Notes: BioMCP auto-downloads ddinter_downloads_code_A.csv, ddinter_downloads_code_B.csv, ddinter_downloads_code_D.csv, ddinter_downloads_code_H.csv, ddinter_downloads_code_L.csv, ddinter_downloads_code_P.csv, ddinter_downloads_code_R.csv, and ddinter_downloads_code_V.csv into BIOMCP_DDINTER_DIR or the default data directory on first use, refreshes stale files after 72 hours, supports explicit refresh via biomcp ddinter sync, and phrases empties as current-bundle misses instead of clinical absence.

gnomAD

• BioMCP surfaces: get gene <symbol> constraint; get variant <id> population
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public browser/API
• License / terms summary: Broad Institute data policies with attribution and service-specific conditions
• Redistribution / reuse summary: querying is open, but users should review the gnomAD policies before bulk reuse or republishing
• Official terms URL: https://gnomad.broadinstitute.org/policies
• Reviewed on: 2026-03-20
• Notes: Open Targets currently reports gnomAD as CC0 for its own ingestion pipeline. BioMCP links to gnomAD's own policy page for user-facing guidance.

Human Protein Atlas

• BioMCP surfaces: get gene <symbol> hpa
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public XML and web resources
• License / terms summary: CC BY-SA 4.0 for copyrightable parts of the database
• Redistribution / reuse summary: reuse is allowed with attribution and ShareAlike; third-party components may impose extra conditions
• Official terms URL: https://www.proteinatlas.org/about/licence
• Reviewed on: 2026-03-20
• Notes: The licence page also requires clear citation for images and specific gene/data pages.

KEGG

• BioMCP surfaces: search pathway; get pathway <id>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open website/API for academic use; commercial use requires a license
• License / terms summary: custom KEGG terms; academic users may freely use the website, non-academic use requires a commercial license
• Redistribution / reuse summary: do not assume commercial redistribution rights; query access does not grant a redistribution license
• Official terms URL: https://www.kegg.jp/kegg/legal.html
• Reviewed on: 2026-03-20
• Notes: KEGG's official legal page was updated on October 1, 2024 and explicitly distinguishes academic from non-academic use.

LitSense2

• BioMCP surfaces: search article -k <query>
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: NCBI/NLM public-domain literature service
• Redistribution / reuse summary: query results are broadly reusable, but preserve article-level provenance and record rights separately
• Official terms URL: https://www.ncbi.nlm.nih.gov/research/litsense2-api/
• Reviewed on: 2026-04-10
• Notes: LitSense2 is only used for keyword-bearing article search paths and contributes semantic-signal metadata rather than acting as the sole article backend.

PharmGKB

• BioMCP surfaces: get pgx <gene_or_drug> annotations
• Integration mode: direct_api
• BioMCP auth: none
• Provider access / registration: open public API
• License / terms summary: ClinPGx API data is CC BY-SA 4.0 and subject to the provider's data usage policy
• Redistribution / reuse summary: reuse is allowed with attribution and ShareAlike; some underlying annotations and external assets may add extra constraints
• Official terms URL: https://api.pharmgkb.org/
• Reviewed on: 2026-03-20
• Notes: PharmGKB has been transitioning to ClinPGx-branded API documentation. BioMCP keeps the public PharmGKB source label because that is the domain vocabulary users recognize.

Indirect-only providers surfaced through aggregators

AlphaFold DB

• BioMCP surfaces: get protein <id> structures
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced through UniProt cross-references; no standalone BioMCP client
• License / terms summary: AlphaFold DB structural predictions are published for broad open use
• Redistribution / reuse summary: reuse is generally open, but preserve model/source provenance and article citations
• Official terms URL: https://alphafold.ebi.ac.uk/faq
• Reviewed on: 2026-03-20
• Notes: BioMCP does not call AlphaFold DB directly. Structure links appear through UniProt cross-references.

Cancer Genome Interpreter

• BioMCP surfaces: get variant <id>
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced indirectly through MyVariant.info payloads
• License / terms summary: custom tool terms
• Redistribution / reuse summary: do not assume commercial reuse rights; the official terms restrict some external and commercial use
• Official terms URL: https://www.cancergenomeinterpreter.org/conditions
• Reviewed on: 2026-03-20
• Notes: There is no standalone CGI source client in BioMCP; provenance appears only when MyVariant includes CGI fields.

ClinVar

• BioMCP surfaces: get variant <id> clinvar
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced indirectly through MyVariant.info payloads
• License / terms summary: NCBI public-domain submission archive
• Redistribution / reuse summary: records are broadly reusable, but preserve accession/provenance and submitter context
• Official terms URL: https://www.ncbi.nlm.nih.gov/clinvar/docs/maintenance_use/
• Reviewed on: 2026-03-20
• Notes: BioMCP does not call ClinVar directly; ClinVar assertions arrive through MyVariant.info.

COSMIC

• BioMCP surfaces: get variant <id> cosmic
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced indirectly through cached MyVariant.info fields; no standalone BioMCP COSMIC client
• License / terms summary: custom COSMIC licensing with commercial restrictions
• Redistribution / reuse summary: direct redistribution and direct integration remain intentionally unsupported without a separate COSMIC license
• Official terms URL: https://www.sanger.ac.uk/legal/cosmic-licensing/
• Reviewed on: 2026-03-20
• Notes: This is the most important indirect-only caution row. BioMCP intentionally does not support direct COSMIC querying because of licensing risk.

Disease Ontology

• BioMCP surfaces: search disease; get disease <id>
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced indirectly through MyDisease.info
• License / terms summary: open disease ontology project
• Redistribution / reuse summary: reuse is generally open; preserve ontology version and source references
• Official terms URL: https://disease-ontology.org/
• Reviewed on: 2026-03-20
• Notes: BioMCP does not maintain a standalone Disease Ontology client.

DrugBank

• BioMCP surfaces: get drug <name> interactions; get drug <name> label
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced indirectly through MyChem.info payloads
• License / terms summary: custom DrugBank terms of use and licensing
• Redistribution / reuse summary: use or redistribution of DrugBank content requires a DrugBank license; do not assume open downstream rights
• Official terms URL: https://trust.drugbank.com/drugbank-trust-center/drugbank-terms-of-service
• Reviewed on: 2026-03-20
• Notes: DrugBank does not have a standalone BioMCP source client. It appears as provenance carried through MyChem.info.

Drugs@FDA

• BioMCP surfaces: get drug <name> approvals
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced through OpenFDA-derived approval fields
• License / terms summary: FDA-origin public information
• Redistribution / reuse summary: approval records are broadly reusable; avoid implying FDA endorsement
• Official terms URL: https://open.fda.gov/apis/drug/drugsfda/
• Reviewed on: 2026-03-20
• Notes: BioMCP documents Drugs@FDA as an indirect provenance label because approval fields arrive through OpenFDA, not a dedicated Drugs@FDA client.

MONDO

• BioMCP surfaces: search disease; discover <query>
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced indirectly through MyDisease.info payloads
• License / terms summary: CC BY 4.0
• Redistribution / reuse summary: reuse is allowed with attribution and ontology version tracking
• Official terms URL: https://mondo.monarchinitiative.org/pages/download/
• Reviewed on: 2026-03-20
• Notes: BioMCP uses MONDO identifiers through MyDisease.info and other aggregators rather than calling MONDO directly.

PDB

• BioMCP surfaces: get protein <id> structures
• Integration mode: indirect_only
• BioMCP auth: not_applicable
• Provider access / registration: surfaced through UniProt cross-references; no standalone BioMCP PDB client
• License / terms summary: PDB archive data is CC0 1.0
• Redistribution / reuse summary: data is broadly reusable; attribution to original structure authors is encouraged
• Official terms URL: https://www.rcsb.org/pages/usage-policy
• Reviewed on: 2026-03-20
• Notes: BioMCP currently exposes PDB identifiers from UniProt rather than querying RCSB PDB directly.

Source notes

• PubMed, PubTator3, Europe PMC, LitSense2, NCBI E-utilities, PMC OA, NCBI ID Converter, Semantic Scholar, and Figshare are separate direct inventory rows inside BioMCP's article stack because search, annotation, identifier, PDF metadata, full-text, and asset-byte responsibilities come from different NCBI/EMBL/AI2/Figshare service surfaces. PMC article HTML is documented as a PMC web fallback in the data-source matrix rather than as a separate source-client row.
• OpenFDA FAERS, OpenFDA label, OpenFDA shortage, and Drugs@FDA are user-facing provenance labels that resolve back to the OpenFDA direct row plus the Drugs@FDA indirect row.
• AlphaFold DB and PDB are indirect-only because BioMCP currently surfaces those structure IDs via UniProt cross-references rather than maintaining standalone source clients.
• COSMIC is indirect-only provenance through MyVariant.info. Direct COSMIC integration is not part of BioMCP's supported source surface because the provider's licensing model creates unacceptable redistribution and deployment risk for an MIT-licensed open tool.