generated: '2026-09-06' method: searched source: >- https://github.com/digital-science/dimensions-analytics-mcp (README, docs/USAGE.md, docs/REFERENCE.md, docs/INSTALLATION.md) + https://docs.dimensions.ai/dsl/mcp.html + probed https://www.dimensions.ai/wp-json/mcp provider: Dimensions providerId: dimensions status: published description: >- Digital Science publishes first-party MCP servers for Dimensions. The flagship is the Dimensions Analytics MCP — an open-source (MIT) local stdio server on npm that exposes 27 tools plus 10 resources over the Dimensions DSL API using the subscriber's own API key. A second, separately reachable remote MCP endpoint runs on the www.dimensions.ai WordPress site and is OAuth-gated. A third server, Dimensions Semantic Search MCP, was announced 2026-08-10 but is request-access only with no public endpoint or package. deployment: mode: both endpoint: https://www.dimensions.ai/wp-json/mcp/mcp-oauth-server install: npx -y @digital-science-dsl/dimensions-analytics-mcp package: https://www.npmjs.com/package/@digital-science-dsl/dimensions-analytics-mcp auth: oauth verified: probed note: >- The two halves are DIFFERENT servers, not two transports of one product. `install` is the Dimensions Analytics MCP (local stdio, Node 20+, API key in DIMENSIONS_API_KEY) — the server documented at docs.dimensions.ai/dsl/mcp.html and the one that reaches research data. `endpoint` is a separate remote MCP server the marketing site runs (WordPress MCP adapter), advertised by RFC 9728 protected-resource metadata and gated by OAuth (scope "mcp", PKCE S256); its tool list is not readable anonymously. A HOSTED Analytics MCP — URL plus bearer token, no Node install — is explicitly "coming soon" and NOT available (docs/INSTALLATION.md, "Hosted Dimensions Analytics MCP (coming soon)"). servers: - name: Dimensions Analytics MCP mode: local-stdio official: true license: MIT repository: https://github.com/digital-science/dimensions-analytics-mcp package: '@digital-science-dsl/dimensions-analytics-mcp' registry: npm version: 1.4.1 published: '2026-09-06' install: npx -y @digital-science-dsl/dimensions-analytics-mcp guided_install: bash -c "$(curl -fsSL https://raw.githubusercontent.com/digital-science/dimensions-analytics-mcp/main/scripts/install.sh)" client_config_key: dimensions runtime: Node.js 20+ auth: api-key auth_detail: >- DIMENSIONS_API_KEY — the same Dimensions Analytics API key used for the DSL API. DIMENSIONS_BASE_URL is required for custom instances (default https://app.dimensions.ai). Subscription-only; MCP access is available to API subscribers under the same Reasonable Use terms. clients_autoconfigured: - Claude Desktop - Cursor - VS Code (Copilot) - Windsurf announced: '2026-08-10' source: https://docs.dimensions.ai/dsl/mcp.html - name: Dimensions marketing-site MCP (WordPress MCP adapter) mode: remote official: true endpoint: https://www.dimensions.ai/wp-json/mcp/mcp-oauth-server transport: streamable-http auth: oauth auth_detail: >- 401 with WWW-Authenticate: Bearer realm="https://www.dimensions.ai", resource_metadata="https://www.dimensions.ai/.well-known/oauth-protected-resource". Authorization server https://www.dimensions.ai, scope "mcp", PKCE S256, token_endpoint_auth_methods_supported ["none"] (public clients), client_id_metadata_document_supported true. probe: url: https://www.dimensions.ai/wp-json/mcp/mcp-oauth-server method: POST tools/list http_status: 401 body: '{"code":"mcp_unauthorized","message":"MCP authentication required."}' fetched: '2026-09-06' tools: [] tools_note: >- Live tool schemas require authenticated introspection. Not enumerated — no tools are asserted for this server. sibling_route: /wp-json/mcp/mcp-adapter-default-server (POST returns 401 rest_forbidden) source: https://www.dimensions.ai/.well-known/oauth-protected-resource - name: Dimensions Semantic Search MCP mode: none official: true status: request-access endpoint: null package: null auth: unknown announced: '2026-08-10' note: >- Announced alongside the Analytics MCP as a concept-aware search layer across 40+ life science domains. No public endpoint, package, or repository was found; the vendor page offers only "Request access". source: https://www.digital-science.com/model-context-protocol-mcp/ tools: - name: search_publications description: Search research articles, preprints, and books with citation metrics - name: search_grants description: Search research grants and funding awards - name: search_researchers description: Search researcher profiles with publication counts and affiliations - name: search_patents description: Search patents with assignee and filing information - name: search_clinical_trials description: Search clinical trials by phase, condition, or status - name: search_datasets description: Search research datasets and repositories - name: search_policy_documents description: Search policy briefs and government documents - name: search_organizations description: Search universities, research institutes, and companies - name: search_reports description: Search technical reports - name: search_source_titles description: Search journals, preprint platforms, book series, and proceedings - name: search_funder_groups description: Search curated funder groups (name to member GRID ids) - name: search_research_org_groups description: Search curated research organization groups (name to member GRID ids) - name: fetch_search_pages description: Paginated / batch search (page, aggregate, JSONL/CSV export) — local stdio only - name: get_by_doi description: Retrieve publication by DOI - name: get_by_pmid description: Retrieve publication by PubMed ID - name: get_by_id description: Retrieve any entity by Dimensions ID - name: construct_profile_url description: Build a canonical Dimensions web profile URL from entity type + ID - name: aggregate_query description: Numeric aggregation (sum/avg/count) over any entity with grouping - name: facet_query description: Facet counts for top values of any field - name: citation_trend description: Citation counts over time for a topic or entity - name: funding_trend description: Funding awarded over time for a topic or entity - name: extract_affiliations description: Resolve organization affiliations to GRID/ROR identifiers - name: extract_grants description: Resolve grant numbers to full grant records - name: similar_documents description: Find publications/grants with similar topics from abstract text (concept-based) - name: execute_dsl description: Execute raw Dimensions DSL queries (classify, extract concepts, KWQ, and more) - name: describe_schema description: Summarize live describe schema (sources, entities, limits) - name: refresh_schema description: Reload describe schema from the API tools_source: >- https://github.com/digital-science/dimensions-analytics-mcp/blob/main/docs/USAGE.md#all-mcp-tools — names and descriptions verbatim from the provider's own table. Input schemas were not read from a live tools/list (the local server requires a subscriber API key), so no parameter shapes are asserted here. resources: - uri: dimensions://schema/summary description: Compact catalog overview — use on cold-start - uri: dimensions://schema description: Full raw describe payload - uri: dimensions://schema/version description: DSL version string - uri: dimensions://schema/limits description: Numeric caps (maxLimit, maxPages, etc.) - uri: dimensions://schema/policy description: Reasonable-use rules and MCP guardrails - uri: dimensions://schema/profile-urls description: Canonical Dimensions web profile URL templates - uri: dimensions://schema/sources/{name} description: Fields, facets, metrics for a source - uri: dimensions://fields/{entityType} description: Filterable/facet fields + alias hints - uri: dimensions://examples description: All curated example queries - uri: dimensions://examples/{source} description: Per-source DSL examples guardrails: rate_limit_per_minute: 30 max_rows_per_call: 1000 default_rows_per_call: 100 max_paginated_records: 50000 facet_buckets: 1000 large_result_warning_threshold: 10000 confirm_large_fetch_required_when: - skip >= 5000 - page >= 5 - limit = 1000 with skip > 0 - maxPages > 5 - planned records > 5000 source: https://docs.dimensions.ai/dsl/usagepolicy.html