Science & Academic Research

Live

Papers, datasets, chemicals, proteins, clinical trials, species and books in one MCP server.

One research server for your assistant, backed by free public sources: scholarly papers and citations (Crossref, Europe PMC), research datasets (DataCite) and institutions (ROR), chemical compounds (PubChem), proteins (UniProt), clinical trial records (ClinicalTrials.gov), species and biodiversity records (GBIF), books (Open Library) and NASA image records. Every answer keeps the source, record ID and dates so you can check it.

Connect this tool ↗Tool reference

Connect / 23 tools

Server URL

streamable http

https://toolcargo.com/mcp/science-research

Claude Code

terminal

claude mcp add --transport http --scope user toolcargo-science-research https://toolcargo.com/mcp/science-research \
  --header "Authorization: Bearer YOUR_API_KEY"

Create a key in your dashboard. Other clients: setup guides.

What it helps you do

Find papers and check their citation details and DOIs.
Look up a chemical compound, a protein or a clinical trial by name or ID.
Find datasets, research institutions, species records and books.
Every result links back to the original public record.

Built for: Researchers, students, analysts and content teams who need sourced facts rather than guesses.

What's inside

One connection gives your assistant all of these. Each has a detailed reference.

Example requests

Literature check

“Find recent papers on GLP-1 and kidney outcomes and give me the citation details.”

Titles, authors, journals, dates and DOIs with source links.

Compound facts

“What is the molecular formula and weight of caffeine? Cite the record.”

PubChem identity and properties with the CID.

Trial status

“Is NCT04368728 still recruiting, and have results been posted?”

Registry status, key dates and results availability.

Explore useful agent workflows ↗

Tools

Tool	What it does	Access
`pubchem_resolve_name` Resolve a complete indexed chemical name	Return up to ten PubChem candidate CIDs for a complete indexed name, with full bounded candidate count and ambiguity. No structure, URL or query-language inputs; a name may resolve to multiple records. One shared-plan call.	Read-only
`pubchem_compound_properties` Research selected calculated compound properties	Verify exact CID/title identity, then inspect molecular formula, calculated molecular weight in g/mol, InChIKey, IUPAC name and summary title. Identity unavailability is not compound absence. Preserve missing fields, raw source types and unknown revisions. No medical, safety or use recommendations. One shared-plan call.	Read-only
`uniprot_gene_search` Find protein records by gene and organism	Search a literal gene symbol and exact organism taxonomy ID in public UniProtKB metadata. Reviewed, unreviewed or any status; default reviewed. Up to ten candidates per explicit filter/release-bound cursor page. Preserve source gene spelling, curation and existence-evidence class; no sequence residues or narratives. One shared-plan call.	Read-only
`uniprot_protein_details` Inspect an exact canonical protein accession	Inspect one exact canonical UniProtKB primary accession: bounded names, genes, exact organism, reviewed status, existence evidence, canonical length, precursor flag, entry/sequence versions and source release. Inactive/redirect/not-found/gone remain source states without biological-absence inference or alias substitution. No sequence residues or medical advice. One shared-plan call.	Read-only
`clinicaltrials_search_studies` Find public study records	Search ClinicalTrials.gov selected public metadata by a plain human condition topic and optional overall status. Explicit filter-bound pages of 1–10 records; no contacts, eligibility, narratives, outcomes or participant data. Fetches the source processing timestamp each operation. One shared-plan call.	Read-only
`clinicaltrials_study_record` Inspect an exact NCT study record	Inspect one canonical eight-digit NCT ID. Preserve source status, independent posted-results availability and actual/estimated month/day dates. Validated source-specific absence only; aliases are not followed. Public registry research, not treatment advice or efficacy/safety findings. One shared-plan call.	Read-only
`gbif_search_species` Search species taxonomy	Search GBIF checklist taxonomies by plain phrase. Return up to 10 taxa per explicit page, scientific names, status, synonym/accepted keys, bounded classification and source links. Checklist taxonKey differs from occurrenceTaxonKey (backbone nubKey). First 1,000 results only. One shared-plan call.	Read-only
`gbif_species_details` Read exact taxon metadata	Inspect an exact positive GBIF taxonKey from search, with identity validation, checklist source, accepted key, classification and occurrenceTaxonKey. No descriptions, images or occurrence records. One shared-plan call.	Read-only
`gbif_occurrence_summary` Summarize indexed occurrence records	Count GBIF records for a backbone occurrenceTaxonKey copied from taxonomy, optionally by uppercase country code and single recorded-event year. Returns at most 10 partial basis-of-record and year facet buckets, without raw records or coordinates. Does not verify taxon existence. Includes child and synonym taxa. Counts are not population size or current presence/absence. One shared-plan call.	Read-only
`search_publications` Search scholarly publications	Search Crossref for up to 10 scholarly metadata records by topic, title or author, optionally within publication years. Returns DOI source links, author names, publication dates and deposited citation counts. Metadata only; no abstracts or full text. One shared-plan call.	Read-only
`publication_details` Look up a publication DOI	Retrieve Crossref bibliographic metadata for one plain DOI, including source link, title, up to 20 author names, journal, publisher, date, licenses and citation/reference counts. Missing metadata is identified. Counts do not measure quality. One shared-plan call.	Read-only
`europe_pmc_search_publications` Search life-sciences publications	Search a literal phrase in Europe PMC, with optional inclusive first-publication year filters. Returns up to 10 bibliographic records with source+ID, dates, publication types, preprint/retraction cautions and an explicit nextCursorMark. No abstracts or full text. One shared-plan call per requested page.	Read-only
`europe_pmc_publication_details` Read publication citation metadata	Read one exact Europe PMC source+ID pair from a search result. Returns bounded citation metadata, source and DOI links, indexed publication types and open-access flag. No abstracts, full text or medical advice. Open access is not a reuse license. One shared-plan call.	Read-only
`datacite_search_datasets` Search dataset DOI metadata	Search public DataCite records classified as Dataset using a literal phrase, optional inclusive publication years and explicit page 1–100. Returns up to 10 DOI-linked metadata records, bounded creator, subject and rights lists. No descriptions or dataset downloads. One shared-plan call.	Read-only
`datacite_dataset_details` Read dataset DOI metadata	Read bounded metadata for one plain DataCite dataset DOI: titles, creators, publisher, year, subjects, deposited rights and related identifiers. Rejects other resource types. No linked files downloaded; verify dataset licensing at the repository. One shared-plan call.	Read-only
`ror_search_organizations` Search research organizations	Search active ROR organization records by name keywords, optionally filtered by country code and organization type. Returns 1–20 projected records from a fixed 20-record upstream page, source IDs, names, status and bounded locations/relationships. Pagination identifies locally omitted records; not a complete registry export. One shared-plan call.	Read-only
`ror_organization_details` Look up a ROR organization	Retrieve one organization by plain nine-character ROR ID, including inactive or withdrawn records. Returns canonical ROR source link, bounded names, city/country locations, websites, organizational relationships and external identifiers. Missing metadata and truncation are explicit. No contact details or coordinates. One shared-plan call.	Read-only
`ol_search_books` Find book works	Human-directed Open Library catalog discovery by plain query, title and/or author. Returns at most 10 work records with distinct indexed first-publication year and edition count. Explicit offset pages only; counts may be approximate. No reader, full text or availability lookup. One shared-plan call.	Read-only
`ol_work_details` Inspect a book work	Inspect one exact OL…W work ID, title, selected subjects, author IDs and catalog first_publish_date. A work groups editions; dates can disagree with the search index. No description, excerpt or linked fetch. One shared-plan call.	Read-only
`ol_work_editions` Compare editions of a work	Read one explicit page of at most 10 editions for an exact OL…W work. Verify each edition's work relationship and compare bounded publisher, date, language and ISBN catalog fields. Offset 0–1000; no crawling or automatic pagination. One shared-plan call.	Read-only
`ol_edition_details` Inspect a book edition	Read bibliographic metadata for one exact OL…M edition ID. Return work/author IDs, publication date, publishers, languages, ISBNs and physical format where recorded. These describe an edition, not all versions. No files, covers, prose or reader access. One shared-plan call.	Read-only
`nasa_search_images` Find space and mission image records	Search NASA Image and Video Library indexed image metadata by human query, optional center and years. At most 10 records per explicit cursor page, capped at page 1000. Recorded credits and inert asset pointers; no image download, visual inspection or reuse clearance. One shared-plan call.	Read-only
`nasa_image_record` Inspect an exact NASA image record	Inspect one case-insensitive exact NASA image ID, preserving canonical source spelling, credits and source dateCreated. Missing licensing stays unassessed. No assets, linked pages, NASA-authored AI output, endorsement or permission claims. One shared-plan call.	Read-only

Requirements

• A ToolCargo account with Site Audit activated (free). No provider accounts or API keys.

Limitations

• Returns what the public source records; it does not judge study quality or give medical advice.
• Each source has its own limits and coverage, described on each tool.
• Single-tool URLs such as /mcp/pubchem remain available for existing keys.

Supported clients

“Tested” means we connected that client to this endpoint and ran a tool call ourselves. Recorded client-specific tests currently cover Site Audit. Other connectors use the documented setup until tested in that client.

Claude Code	API key header	Documented, not yet tested	This endpoint has not yet been tested in this client. See the setup guide.
Claude (claude.ai)	OAuth sign-in	Documented, not yet tested	This endpoint has not yet been tested in this client. See the setup guide.
Cursor	API key header	Documented, not yet tested	This endpoint has not yet been tested in this client. See the setup guide.
VS Code (Copilot agent mode)	API key header	Documented, not yet tested	This endpoint has not yet been tested in this client. See the setup guide.
ChatGPT (developer mode)	OAuth sign-in	Documented, not yet tested	This endpoint has not yet been tested in this client. See the setup guide.
MCP Inspector	API key header	Documented, not yet tested	This endpoint has not yet been tested in this client. See the setup guide.

Practical workflows

Economy & Public Data LiveEconomic indicators, exchange rates, weather, tides, earthquakes and holiday calendars in one MCP server.

Pricing

Included with Site Audit. Each successful lookup uses one shared-plan call. The public sources are free; ToolCargo fees cover hosting and workflows.

Free

Try hosted tools with a shared monthly allowance.

50 shared calls/month

Pro

$12/mo

Proposed larger shared allowance; see checkout status on the pricing page.

2,000 shared calls/month

Proposed test price. See the pricing page for checkout status.

Full pricing and billing details