pubmed-multi-source-search

Multi-source comprehensive search

Description

The currently public multi-source search entry point is unified_search. It automatically routes across PubMed, OpenAlex, Semantic Scholar, Europe PMC, and CrossRef, and performs integration and deduplication within a single search.

Trigger conditions

"cross-database search"
"comprehensive search"
"don't only look at PubMed"
Mentions OpenAlex, Semantic Scholar, Europe PMC, preprint
Need to supplement coverage across sources

Core principles

The new workflow no longer directly depends on multiple source-specific MCP tools. For most literature search scenarios, prioritize using unified_search, rather than the old source-specific tool names.

Core tools

unified_search(
    query="machine learning drug discovery",
    sources="pubmed,openalex,semantic_scholar,europe_pmc",
    limit=25,
    ranking="balanced",
    output_format="json"
)

`sources` optional values

pubmed
openalex
semantic_scholar
europe_pmc
crossref

If not specified, the system will automatically select sources.

Common usages

1. Broad coverage

unified_search(
    query="machine learning drug discovery",
    limit=30,
    ranking="balanced"
)

2. Only look at core biomedical sources

unified_search(
    query="sepsis biomarkers",
    sources="pubmed,europe_pmc",
    limit=25,
    ranking="quality"
)

3. Include preprints

unified_search(
    query="COVID-19 vaccine efficacy",
    sources="pubmed,europe_pmc,openalex",
    options="preprints",
    limit=30,
    ranking="recency"
)

4. Look at high-impact cross-disciplinary literature

unified_search(
    query="foundation models pathology",
    sources="pubmed,openalex,semantic_scholar",
    limit=30,
    ranking="impact"
)

5. Programmatic post-processing

unified_search(
    query="CRISPR gene therapy",
    sources="pubmed,openalex,semantic_scholar,europe_pmc",
    limit=20,
    output_format="json"
)

Recommended workflows

Scenario 1: Find complete coverage across sources

unified_search(
    query="remimazolam sedation",
    sources="pubmed,europe_pmc,openalex,semantic_scholar",
    limit=30,
    ranking="balanced"
)

Scenario 2: First find literature, then supplement full text

unified_search(
    query="machine learning radiology",
    limit=20,
    output_format="json"
)

# 對選中的 PMID / DOI 進一步抓全文
get_fulltext(pmid="12345678", extended_sources=True)

Scenario 3: Explore after cross-source search

unified_search(
    query="CAR-T lymphoma",
    limit=15,
    ranking="impact"
)

find_related_articles(pmid="12345678")
find_citing_articles(pmid="12345678")
get_citation_metrics(pmids="12345678,23456789")

When to adjust ranking

balanced: default, general exploration
impact: first find high-impact representative works
recency: first look at the latest research and updates
quality: favor high evidence levels and more reliable research

When to use options

options="preprints, shallow"

Common combinations

preprints: want to supplement the latest not yet formally published research
shallow: only want to quickly scan through
no_relax: do not want automatic relaxation when there are zero results
no_analysis: programmatic output is cleaner

Scenarios not for this skill

Need to do complete MeSH / synonym expansion: switch to pubmed-systematic-search
It is a clinical comparison question: switch to pubmed-pico-search
Want to find images: use search_biomedical_images
Want to retrieve full text: use pubmed-fulltext-access

Final principle

For textual literature search in the public MCP interface, prioritize unified_search , and no longer split it into source-specific search workflows.