DataCite MCP Server

DataCite

DOI registration and scholarly metadata API. Search for research works by keyword, fetch detailed metadata by DOI, and explore citation/relationship graphs. Public API, no authentication required.

Entity Model

• DOIs — persistent identifiers for scholarly works (papers, datasets, software, etc.). Each DOI has rich metadata: title, authors, publication year, resource type, publisher, descriptions, subjects, and funding references.
• Related Works — citation and relationship graph for a given DOI. Shows works that cite, are cited by, reference, or are related to the target work. Each related work has its own DOI and relationship type.

Workflows

• Literature search: search DOIs by keyword query → review results with titles, authors, and publication years → fetch full metadata for promising hits.
• Citation exploration: search for a known work → get its DOI → fetch related works to explore the citation network (what it cites, what cites it).
• Dataset discovery: search with resource type filters to find datasets, software, or other non-article scholarly outputs.

Gotchas

• No auth: completely public API; no credentials required.
• Rate limit: ~100 requests per minute. Keep page[size] small (3-10) to stay within limits.
• DOI format: always use the full DOI string (e.g., 10.1234/example), not just the suffix or a URL.
• JSON:API format: responses follow the JSON:API specification. Data is in data array, each item has id (the DOI), type, and attributes with the metadata fields.
• Pagination: uses page[size] and page[number] parameters (bracket syntax).