Digital Science today launched two new Model Context Protocol (MCP) servers for its Dimensions research platform, giving enterprise AI agents live access to one of the world's largest interconnected research databases. The integrations address a core problem for R&D teams: AI agents drawing on stale or unverifiable data, producing confident-sounding but unsourced answers.
The launch comprises two purpose-built MCP servers. Dimensions Semantic Search MCP is a concept-aware search layer designed for precision retrieval across 40-plus life science domains. Dimensions Analytics MCP connects AI agents to more than 430 million interconnected records spanning publications, grants, patents, clinical trials, datasets, and policy documents. Both servers are available immediately to existing Dimensions API customers at no additional license cost.
For research-intensive organizations, the typical workaround for unreliable AI outputs is manual data aggregation from fragmented sources-a process that is time-intensive, error-prone, and difficult to scale. The Dimensions MCP servers aim to replace that workflow by letting teams automate research discovery, funding intelligence, and competitive landscape analysis from within the tools they already use.
How semantic search changes scientific retrieval
Traditional keyword search finds only documents containing the exact term used in the query. Dimensions Semantic Search MCP works differently: it identifies the underlying scientific concept and uses domain ontologies to surface related substances, diseases, and compounds automatically.
Peter Haase, VP Knowledge Graph Technologies at Digital Science, explained the practical difference: "A traditional keyword search for 'PFAS' only finds documents containing that exact term. Semantic search, by contrast, identifies the underlying scientific concept and uses domain ontologies to recognize the substances that belong to that concept, such as PFOS, PFOA, PFHxS, and others. It enables researchers to find scientific evidence based on concepts rather than keywords, bringing search closer to the way domain experts think about a subject."
The search layer covers publications, patents, clinical trials, and drug labels within a single interface, with precision retrieval at the section level. It works across every major AI platform without specialist setup.
For professionals working in drug discovery, medical affairs, biotechnology, or regulatory intelligence, that shift-from keywords to meaning-could mean finding relevant evidence that a keyword-only search would miss entirely.
Connecting AI to live, structured research data
The Analytics MCP server gives AI assistants access to Dimensions' linked view of global research activity. Teams can map competitive landscapes across therapeutic areas, geographies, or technology domains; profile research organizations and investigators by aggregating publications, grants, and funding relationships automatically; and track funding trends identifying top funders, award sizes, and activity across topics or institutions.
Sebastian Schmidt, EVP Enterprise at Digital Science, said: "Research-intensive organizations have invested significantly in AI-the models, the workflows, the infrastructure. What they need is an authoritative bridge between AI and research intelligence. Our new Dimensions MCP integrations do exactly that."
The MCP standard is compatible with every major AI platform, meaning teams can use the new integrations from AI for Science & Research tools they already have in place. Used together, the two MCP servers give AI agents the depth to find precise evidence and the breadth to map the research landscape around it-all through a single Dimensions API connection.
Why this matters for science and research professionals
For R&D teams making high-stakes decisions about drug targets, funding strategies, or competitive positioning, the standard AI approach-querying models trained on static datasets months or years old-introduces material risk. Dimensions MCP servers address that by running queries against live, structured data that is linked at the document, person, organization, and funding-source level. That means a query about emerging work in a therapeutic area draws on the same record set a human analyst would use, not a probabilistic guess based on training data. For professionals whose job depends on the accuracy and timeliness of research intelligence, that shift from hallucination risk to verifiable data access is the tangible difference.
More details on the integrations are available at Dimensions MCP integrations page.
Your membership also unlocks:
