PubTator3 through GeneFoundry
Search biomedical literature and inspect machine-annotated entities and relations from PubTator3.
- Research area
- Literature
- Namespace
pubtator- Listed tools
- 35
- Upstream source
- PubTator3
Research tasks
- Find PubMed literature for a research question
- Follow an entity annotation back to the article
Start with tool discovery
The catalog lists pubtator_search_literature as a representative tool. After connecting your client, inspect the tools and input requirements returned by the service before submitting a query.
Identifiers to prepare
- PubMed search text
- PMID
- PMCID
- PubTator entity identifiers
All available tools (35)
The router surfaces these tools namespaced under pubtator_*:
get_server_capabilitiesUse this when a client needs supported tools, transports, formats, and limitations. Do not use this for task-specific workflow guidance; use workflow_help. Next: workflow_help.
Use this when a client needs supported tools, transports, formats, and limitations. Do not use this for task-specific workflow guidance; use workflow_help. Next: workflow_help.
Input parameters
details(array) — Optional capability sections to expand (e.g. 'tools', 'workflow_help'); omit for the default summary.
workflow_helpUse this when a fresh context needs the canonical PubTator-Link research workflow.
Use this when a fresh context needs the canonical PubTator-Link research workflow.
Input parameters
task(string) — Workflow to describe; defaults to the canonical 'clinical_genetics_review' pipeline.
search_literatureUse this when a user needs PubMed literature search through PubTator3. Supports flat filters, section filters, and coverage='preflight'. If preflight_error_code is coverage_preflight_internal_error, retryable=false means continue with results or inspect diagnostics.
Use this when a user needs PubMed literature search through PubTator3. Supports flat filters, section filters, and coverage='preflight'. If preflight_error_code is coverage_preflight_internal_error, retryable=false means continue with results or inspect diagnostics.
Input parameters
coverage(string) — 'none' (default) or 'preflight' to attach source-coverage hints.entity_ids(array) — Restrict to articles mentioning these PubTator entity IDs (resolve them first with search_biomedical_entities), AND-combined with the query.filters(string) — Advanced PubTator3 filter as a JSON object string, e.g. '{"journal":["Nature"]}'. Prefer the flat publication_types / year_min / year_max parameters; do not also set a JSON `type`/`year` key.guideline_boost(boolean) — Boost guideline / systematic-review / consensus articles in ranking.include_citations(string) — Citation rendering per hit: 'none' (default), 'nlm', 'bibtex', or 'both'.include_meta(boolean) — Include the _meta orientation block (next_commands, provenance).limit(integer) — Maximum hits to return on this page.metadata(string) — Metadata enrichment per hit: 'none', 'basic' (default), 'with_abstract', or 'full'.page(integer) — 1-based page number for paging beyond `limit`.publication_types(array) — Restrict to these PubMed publication types (case-sensitive, Title-Case), AND-combined with the query.response_mode(string) — Payload verbosity: 'compact' (default, LLM-friendly), 'standard', or 'full'.sections(array) — Restrict the text match to these article sections (lowercase, case-sensitive).sort(string) — Sort order. Accepts 'date desc' (newest first), 'score desc' (relevance, default), or '_id desc'. Synonyms such as 'date' or 'relevance' are normalized; PubTator3 sorts descending only.text(string) [required] — Free-text PubMed/PubTator3 query: entity names, gene symbols, HGVS, or a natural-language topic. PubTator3 matches across title, abstract, and, for open-access articles, full text.text_hl_format(string) — Match-highlight rendering: 'none', 'plain' (default), or 'annotated'.year_max(integer) — Latest publication year, inclusive.year_min(integer) — Earliest publication year, inclusive.
search_guidelinesUse this when a user needs guideline, recommendation, consensus, or systematic review papers for a biomedical research question. Wraps search_literature with guideline/systematic-review filters and guideline boosting; not an independent guideline database.
Use this when a user needs guideline, recommendation, consensus, or systematic review papers for a biomedical research question. Wraps search_literature with guideline/systematic-review filters and guideline boosting; not an independent guideline database.
Input parameters
coverage(string) — 'preflight' (default) or 'none' to skip source-coverage hints.entity_ids(array) — Restrict to articles mentioning these PubTator entity IDs (resolve them first with search_biomedical_entities).limit(integer) — Maximum hits to return on this page.page(integer) — 1-based page number for paging beyond `limit`.sections(array) — Restrict the text match to these article sections (lowercase).text(string) [required] — Free-text research question to find guideline, recommendation, consensus, or systematic-review articles for.year_max(integer) — Latest publication year, inclusive.year_min(integer) — Earliest publication year, inclusive.
search_biomedical_entitiesUse this when a user needs canonical PubTator biomedical entity IDs for genes, diseases, chemicals, species, variants, or cell lines.
Use this when a user needs canonical PubTator biomedical entity IDs for genes, diseases, chemicals, species, variants, or cell lines.
Input parameters
concept(string) — Restrict autocomplete to one PubTator concept type; omit for all types.limit(integer) — Maximum entity candidates to return.query(string) [required] — Entity name or fragment to resolve, e.g. a gene symbol or disease name.
find_entity_relationsUse this when a user has a PubTator entity ID and needs literature-derived related entities to expand a corpus. Do not use this for canonical entity lookup; use search_biomedical_entities. Next: search_literature.
Use this when a user has a PubTator entity ID and needs literature-derived related entities to expand a corpus. Do not use this for canonical entity lookup; use search_biomedical_entities. Next: search_literature.
Input parameters
entity_id(string) [required] — PubTator entity ID to expand from, e.g. @CHEMICAL_remdesivir (resolve names with search_biomedical_entities first).limit(integer) — Maximum related entities to return.max_response_chars(integer) — Soft character budget for the response body.relation_type(string) — Optional relation type to keep, e.g. 'treat', 'cause', 'associate'; omit for all relation types.response_mode(string) — Payload verbosity: 'compact' (default), 'standard', or 'full'.target_entity_type(string) — Optional target concept type to keep, e.g. 'Disease' or 'Chemical'; omit for all target types.
get_variant_evidenceUse this when a user needs source-attributed variant records and literature evidence for a gene and variant. Does not compute clinical classification.
Use this when a user needs source-attributed variant records and literature evidence for a gene and variant. Does not compute clinical classification.
Input parameters
condition(string) — Optional condition/phenotype to scope the evidence.gene(string) [required] — HGNC gene symbol the variant sits in.include_citations(boolean) — Attach formatted citations for the supporting PMIDs.max_literature_pmids(integer) — Maximum supporting literature PMIDs to attach.protein(string) — Protein-level change, e.g. p.Glu23fs, if known.sources(array) — Evidence sources to include; omit for all.variant(string) [required] — cDNA/HGVS or rsID variant string. Required; use `protein`/`condition` to refine.
suggest_corpusUse this when a user needs a compact, review-feeding PMID corpus for a research question. Returns candidate PMIDs, roles, coverage hints, metadata, and next commands.
Use this when a user needs a compact, review-feeding PMID corpus for a research question. Returns candidate PMIDs, roles, coverage hints, metadata, and next commands.
Input parameters
entity_ids(array) — Optional PubTator entity IDs to anchor the corpus on.include_metadata(boolean) — Attach per-PMID citation metadata to each candidate.max_pmids(integer) — Maximum candidate PMIDs to return.must_include_pmids(array) — PMIDs that MUST appear in the returned corpus.prefer_guidelines(boolean) — Bias selection toward guideline / review articles.question(string) [required] — Research question to assemble a compact candidate PMID corpus for.
convert_article_idsUse this when a user provides article identifiers such as PMIDs, PMCIDs, or DOIs and needs normalized candidate PMIDs for research workflows.
Use this when a user provides article identifiers such as PMIDs, PMCIDs, or DOIs and needs normalized candidate PMIDs for research workflows.
Input parameters
ids(array) [required] — Article identifiers (PMIDs, PMCIDs, or DOIs) to normalize to PMIDs.source(string) — Identifier kind: 'auto' (default, detect), 'pmid', 'pmcid', or 'doi'.
get_meshUse this when a user needs MeSH descriptors and candidate PubMed search terms for a biomedical research query.
Use this when a user needs MeSH descriptors and candidate PubMed search terms for a biomedical research query.
Input parameters
exact(boolean) — Require an exact descriptor match instead of prefix/fuzzy.limit(integer) — Maximum MeSH descriptors to return.query(string) [required] — Term to resolve to MeSH descriptors and candidate search terms.
get_citationUse this when a user provides free-text citations and needs candidate PMIDs for research evidence gathering.
Use this when a user provides free-text citations and needs candidate PMIDs for research evidence gathering.
Input parameters
citations(array) [required] — Free-text citation strings to resolve to candidate PMIDs.
find_related_articlesUse this when a user has seed PMIDs and needs similar, cited-by, or reference-linked articles to expand a research corpus.
Use this when a user has seed PMIDs and needs similar, cited-by, or reference-linked articles to expand a research corpus.
Input parameters
limit(integer) — Maximum related articles to return.mode(string) — Relation to follow: 'similar' (default), 'cited_by', or 'references'.pmids(array) [required] — Seed PMIDs to expand from.
diagnosticsUse this when a client needs PubTator-Link subsystem status and recovery commands.
Use this when a client needs PubTator-Link subsystem status and recovery commands.
get_publication_annotationsUse this when a user provides PubMed IDs and needs raw PubTator BioC annotation export. Do not use this for compact grounded answers; use get_publication_passages. Next: get_publication_passages.
Use this when a user provides PubMed IDs and needs raw PubTator BioC annotation export. Do not use this for compact grounded answers; use get_publication_passages. Next: get_publication_passages.
Input parameters
format(string) — Export serialization: 'biocjson' (default), 'biocxml', or 'pubtator'.full(boolean) — Request full-text annotations where available (else abstract).pmids(array) [required] — PubMed IDs to export raw PubTator BioC annotations for.
build_topic_literature_mapUse this when a user needs a bounded topic-level literature map from a topic query, optionally seeded with PMIDs. Returns response_size_class. response_mode='compact' is the MCP default for LLM candidate selection; full can be large and is for explicit debug graph inspection. Next: get_publication_passages.
Use this when a user needs a bounded topic-level literature map from a topic query, optionally seeded with PMIDs. Returns response_size_class. response_mode='compact' is the MCP default for LLM candidate selection; full can be large and is for explicit debug graph inspection. Next: get_publication_passages.
Input parameters
bias_toward(array) — Bias ranking toward these evidence flavors.citation_graph_timeout_ms(integer) — Per-step timeout for citation-graph lookups.expand_query_seeds(boolean) — Seed the map from a query search when no PMIDs are given.include_authors(boolean) — Include author lists on graph nodes.include_citations(boolean) — Include citation edges in the graph.include_demoted(boolean) — Include demoted (lower-ranked) candidates.include_meta(boolean) — Include the _meta orientation block.include_pubtator_entities(boolean) — Attach PubTator entity annotations to nodes.include_related_candidates(boolean) — Include related-evidence candidates in the result.max_candidates(integer) — Maximum ranked candidate papers.max_demoted(integer) — Maximum demoted candidates to include.max_graph_edges(integer) — Maximum edges in the returned graph.max_graph_nodes(integer) — Maximum nodes in the returned graph.max_neighbors_per_paper(integer) — Maximum neighbors per seed paper.max_seed_papers(integer) — Maximum seed papers to expand.metadata_backfill_timeout_ms(integer) — Per-step timeout for metadata backfill.partial_ok(boolean) — Return a partial map if a sub-step times out.pmids(array) — Optional seed PMIDs to anchor the map on.prefer_full_text(boolean) — Prefer open-access full-text candidates.query(string) [required] — Topic or research question to build the literature map around.related_evidence_timeout_ms(integer) — Per-step timeout for related-evidence lookups.response_mode(string) — Payload shape: 'compact' (default), 'nodes_edges', or 'full'.timeout_ms(integer) — Overall soft timeout in milliseconds.year_max(integer) — Latest publication year, inclusive.year_min(integer) — Earliest publication year, inclusive.
get_publication_passagesUse this when a user needs compact citable publication passages from PMIDs without raw BioC. For article-local answering, use mode='full_abstract' first; it returns all title/abstract passages without truncating structured abstracts. If full=True returns only abstracts, inspect coverage_by_pmid and answer from available evidence. Do not use for prepared review RAG; use get_review_context_batch.
Use this when a user needs compact citable publication passages from PMIDs without raw BioC. For article-local answering, use mode='full_abstract' first; it returns all title/abstract passages without truncating structured abstracts. If full=True returns only abstracts, inspect coverage_by_pmid and answer from available evidence. Do not use for prepared review RAG; use get_review_context_batch.
Input parameters
dry_run(boolean) — Return a size/coverage estimate without passage text.full(boolean) — Prefer full-text passages where the article is open-access.include_references(boolean) — Include reference-list passages.include_tables(boolean) — Include table passages.max_chars(integer) — Soft total character budget for passages.max_passages_per_pmid(integer) — Maximum passages returned per PMID.mode(string) — Passage selection: 'compact_passages' (default), 'full_abstract' (all title/abstract passages), 'abstracts', or 'section_text'.pmids(array) [required] — PubMed IDs to fetch compact citable passages for.sections(array) — Restrict to these BioC section labels (case-insensitive); omit for all sections. An article that lacks a requested section simply contributes no passages for it.verbosity(string) — Field verbosity: 'lean', 'standard' (default), or 'full'.
get_publication_metadataUse this when a user needs citation-grade metadata for known PMIDs. Do not use this for article text or annotations; use get_publication_passages. Next: get_publication_passages.
Use this when a user needs citation-grade metadata for known PMIDs. Do not use this for article text or annotations; use get_publication_passages. Next: get_publication_passages.
Input parameters
include_citations(string) — Citation rendering: 'none', 'nlm', 'bibtex', or 'both' (default).include_coverage(boolean) — Include per-PMID source-coverage hints.include_mesh(boolean) — Include MeSH descriptors.include_publication_types(boolean) — Include PubMed publication types.pmids(array) [required] — PubMed IDs to fetch citation-grade metadata for.
get_publication_citation_graphUse this when a user needs reference or cited-by neighbors for one publication. Returns response_size_class. response_mode='compact' is the MCP default for LLM candidate selection; full can be large and is for explicit debug graph inspection. Next: get_publication_passages.
Use this when a user needs reference or cited-by neighbors for one publication. Returns response_size_class. response_mode='compact' is the MCP default for LLM candidate selection; full can be large and is for explicit debug graph inspection. Next: get_publication_passages.
Input parameters
direction(string) — Which neighbors: 'references', 'cited_by', or 'both' (default).include_open_access_status(boolean) — Include open-access status per neighbor.include_provider_status(boolean) — Include per-provider availability status.max_reference_resolution(integer) — Maximum reference DOIs to resolve.max_results(integer) — Maximum neighbor publications to return.pmid(string) [required] — PubMed ID of the publication whose citation neighbors are wanted.resolve_metadata(boolean) — Resolve title/author metadata for neighbor PMIDs.resolve_reference_pmids(boolean) — Resolve DOIs in the reference list back to PMIDs.response_mode(string) — Payload shape: 'compact' (default), 'nodes_edges', or 'full'.
find_related_evidence_candidatesUse this when a user has one PMID and needs related full-text-preferred candidates. Returns response_size_class. response_mode='compact' is the MCP default for LLM candidate selection; full can be large and is for explicit debug graph inspection. Next: get_publication_passages.
Use this when a user has one PMID and needs related full-text-preferred candidates. Returns response_size_class. response_mode='compact' is the MCP default for LLM candidate selection; full can be large and is for explicit debug graph inspection. Next: get_publication_passages.
Input parameters
citation_graph_timeout_ms(integer) — Per-step timeout for citation-graph lookups.include_citation_neighbors(boolean) — Include citation-graph neighbors.include_meta(boolean) — Include the _meta orientation block.include_pubtator_search(boolean) — Include PubTator entity-search neighbors.max_results(integer) — Maximum candidate publications to return.metadata_timeout_ms(integer) — Per-step timeout for metadata resolution.pmid(string) [required] — Seed PubMed ID to find related evidence candidates for.prefer_full_text(boolean) — Prefer open-access full-text candidates.publication_types(array) — Restrict candidates to these PubMed publication types.response_mode(string) — Payload shape: 'compact' (default), 'nodes_edges', or 'full'.year_max(integer) — Latest publication year, inclusive.year_min(integer) — Earliest publication year, inclusive.
estimate_publication_contextUse this when a user needs to estimate passage count and context size before fetching publication passages. Do not use this for text retrieval; use get_publication_passages. Next: get_publication_passages.
Use this when a user needs to estimate passage count and context size before fetching publication passages. Do not use this for text retrieval; use get_publication_passages. Next: get_publication_passages.
Input parameters
full(boolean) — Estimate under full-text retrieval where available.include_references(boolean) — Count reference-list passages in the estimate.include_tables(boolean) — Count table passages in the estimate.max_passages_per_pmid(integer) — Maximum passages per PMID to assume.mode(string) — Passage selection to estimate under: 'compact_passages' (default), 'full_abstract', 'abstracts', or 'section_text'.pmids(array) [required] — PubMed IDs to estimate passage count and context size for.sections(array) — Restrict to these BioC section labels (case-insensitive); omit for all sections. An article that lacks a requested section simply contributes no passages for it.
get_pmc_annotationsUse this when a user provides PMC IDs and needs raw PubTator full-text BioC annotation export. Do not use this for compact grounded answers; use get_publication_passages. Next: get_publication_passages.
Use this when a user provides PMC IDs and needs raw PubTator full-text BioC annotation export. Do not use this for compact grounded answers; use get_publication_passages. Next: get_publication_passages.
Input parameters
format(string) — Export serialization: 'biocjson' (default) or 'biocxml'.pmcids(array) [required] — PMC IDs to export raw PubTator full-text BioC annotations for.
get_text_annotation_resultsUse this when a user has a PubTator text annotation session ID and needs its results. Do not use this for entity lookup from names; use search_biomedical_entities. Next: search_biomedical_entities.
Use this when a user has a PubTator text annotation session ID and needs its results. Do not use this for entity lookup from names; use search_biomedical_entities. Next: search_biomedical_entities.
Input parameters
session_id(string) [required] — PubTator text-annotation session ID returned by a prior request.
list_review_indexesUse this when a user needs persisted review indexes with preparation status, source counts, passage counts, and approximate storage size.
Use this when a user needs persisted review indexes with preparation status, source counts, passage counts, and approximate storage size.
Input parameters
limit(integer) — Maximum review indexes to return.offset(integer) — Zero-based row offset for pagination.
get_review_index_summaryUse this when a user needs one persisted review index summary without loading passage samples.
Use this when a user needs one persisted review index summary without loading passage samples.
Input parameters
review_id(string) [required] — Identifier of the persisted review index to summarize.
inspect_review_indexUse this when a user needs to inspect indexed PMIDs, sections, passage counts, and failures for a review_id, including source coverage.
Use this when a user needs to inspect indexed PMIDs, sections, passage counts, and failures for a review_id, including source coverage.
Input parameters
cursor(string) — Opaque pagination cursor from a prior page.include_metadata(boolean) — Include per-PMID citation metadata.include_passage_samples(boolean) — Include a few sample passages per PMID.limit(integer) — Maximum indexed PMIDs to page over.metadata(string) — Metadata depth when included: 'basic' (default) or 'full'.min_sample_chars(integer) — Minimum characters for a sampled passage.pmids(array) — Restrict the inspection to these indexed PMIDs.response_mode(string) — Payload verbosity: 'compact' (default) or 'full'.review_id(string) [required] — Review index to inspect.sample_per_pmid(integer) — Sample passages to include per PMID.sample_section_policy(string) — Sampling order: 'evidence_first' (default) or 'original_order'.session_id(string) — Optional staged session to scope the inspection to.
list_evidence_certaintyUse this when a user needs user-supplied evidence certainty judgments for a review.
Use this when a user needs user-supplied evidence certainty judgments for a review.
Input parameters
review_id(string) [required] — Review whose stored certainty judgments to list.
get_evidence_certaintyUse this when a user needs one user-supplied evidence certainty judgment.
Use this when a user needs one user-supplied evidence certainty judgment.
Input parameters
certainty_id(string) [required] — Identifier of the stored certainty judgment to fetch.review_id(string) [required] — Review the certainty judgment belongs to.
preflight_review_sourcesUse this when a user needs PMID source coverage, PMC fallback availability, and likely full-text versus abstract-only retrieval before indexing review evidence.
Use this when a user needs PMID source coverage, PMC fallback availability, and likely full-text versus abstract-only retrieval before indexing review evidence.
Input parameters
pmid(string) — Single-PMID convenience alias, merged with `pmids`.pmids(array) [required] — PubMed IDs to check source coverage and full-text availability for.
get_research_session_statusUse this when a user needs staged candidate, coverage, and preparation status for a research session.
Use this when a user needs staged candidate, coverage, and preparation status for a research session.
Input parameters
review_id(string) — Optional review index the session belongs to.session_id(string) [required] — Staged research session to report status for.
list_research_sessionsUse this when a user needs staged research sessions for orientation or one review ID.
Use this when a user needs staged research sessions for orientation or one review ID.
Input parameters
cursor(string) — Opaque cursor returned by a previous session-list page.limit(integer) — Maximum compact session summaries to return.review_id(string) — Optional review index to list sessions for; omit for recent global sessions.
get_review_passages_by_idUse this when a user needs exact prepared review passages by stable passage IDs from prior context packs or audit bundles. This only reads the review index and does not call upstream APIs.
Use this when a user needs exact prepared review passages by stable passage IDs from prior context packs or audit bundles. This only reads the review index and does not call upstream APIs.
Input parameters
max_chars_per_passage(integer) — Character cap per returned passage.passage_ids(array) [required] — Stable prepared passage IDs to fetch verbatim.review_id(string) [required] — Review index the passages belong to.session_id(string) — Optional staged session to scope the lookup to.
get_review_audit_trailUse this when a user needs a copy-ready audit block for selected prepared review passage IDs or the latest recorded audit passages without calling upstream APIs.
Use this when a user needs a copy-ready audit block for selected prepared review passage IDs or the latest recorded audit passages without calling upstream APIs.
Input parameters
max_chars_per_passage(integer) — Character cap per audited passage.passage_ids(array) — Specific prepared passage IDs to audit; omit for the latest recorded set.review_id(string) [required] — Review index to build the audit block for.session_id(string) — Optional staged session to scope the audit to.
get_neighboring_review_passagesUse this when a user needs prepared review passages near a cited stable passage ID for local context expansion. This only reads the review index and does not call upstream APIs.
Use this when a user needs prepared review passages near a cited stable passage ID for local context expansion. This only reads the review index and does not call upstream APIs.
Input parameters
after(integer) — Neighboring passages to include after the anchor.before(integer) — Neighboring passages to include before the anchor.max_chars_per_passage(integer) — Character cap per returned passage.passage_id(string) [required] — Anchor passage ID to expand context around.review_id(string) [required] — Review index the passage belongs to.same_section(boolean) — Restrict neighbors to the anchor's section.session_id(string) — Optional staged session to scope the lookup to.
get_review_contextUse this when a review needs compact citable context from prepared review passages instead of raw BioC export. Use a short keyword query and PMID filters. If zero passages are returned, simplify the query, inspect the review index, or fall back to get_publication_annotations.
Use this when a review needs compact citable context from prepared review passages instead of raw BioC export. Use a short keyword query and PMID filters. If zero passages are returned, simplify the query, inspect the review index, or fall back to get_publication_annotations.
Input parameters
allow_truncated_passages(boolean) — Allow per-passage truncation to fit the char budget.entity_ids(array) — Restrict retrieval to these PubTator entity IDs.include_diagnostics(boolean) — Include retrieval diagnostics in the response.include_meta(boolean) — Include the _meta orientation block.include_references(boolean) — Include reference-list passages.include_resolver_trace(boolean) — Include the source-resolver trace for auditing.include_tables(boolean) — Include table passages.max_chars(integer) — Soft total character budget for passages.max_chars_per_passage(integer) — Character cap per returned passage.max_passages(integer) — Maximum passages to return.pmids(array) — Restrict retrieval to these PMIDs.question(string) [required] — Short keyword retrieval question or query.review_id(string) [required] — Prepared review index to retrieve citable context from.section_policy(string) — Passage ordering: 'evidence_first' (default) or 'original_order'.sections(array) — Restrict retrieval to these article sections.session_id(string) — Optional staged session to scope retrieval to.table_mode(string) — Table rendering: 'off', 'preview' (default), or 'full'.
get_review_context_batchUse this when a user wants multiple short review retrieval query variants in one call. Default compact mode uses query_fair budgeting, merged passages, per-query summaries, and next_steps for zero-result queries. Use response_mode="quotes" for short citable snippets or dry_run for diagnostics without passage text.
Use this when a user wants multiple short review retrieval query variants in one call. Default compact mode uses query_fair budgeting, merged passages, per-query summaries, and next_steps for zero-result queries. Use response_mode="quotes" for short citable snippets or dry_run for diagnostics without passage text.
Input parameters
allow_truncated_passages(boolean) — Allow per-passage truncation to fit the char budget.budget_strategy(string) — Passage budget split (default 'query_fair').deduplicate_passages(boolean) — Merge duplicate passages across query variants.dry_run(boolean) — Return budgeting diagnostics without passage text.entity_ids(array) — Restrict retrieval to these PubTator entity IDs.include_diagnostics(boolean) — Include retrieval diagnostics in the response.include_meta(boolean) — Include the _meta orientation block.include_references(boolean) — Include reference-list passages.include_resolver_trace(boolean) — Include the source-resolver trace for auditing.include_tables(boolean) — Include table passages.max_chars(integer) — Optional soft total character budget.max_chars_per_passage(integer) — Character cap per returned passage.max_passages_per_query(integer) — Maximum passages per query variant.max_response_chars(any) — Response character budget: 'auto' (default) or an integer cap.max_total_passages(integer) — Maximum passages across all queries.min_passages_per_pmid(integer) — Guaranteed minimum passages per PMID.min_passages_per_source(integer) — Guaranteed minimum passages per source.pmids(array) — Restrict retrieval to these PMIDs.prioritize_pmids(array) — PMIDs to prioritize in budgeting.queries(array) [required] — Short keyword query variants to retrieve context for in one call.response_mode(string) — Payload shape (default 'compact'); 'quotes' for short snippets.review_id(string) [required] — Prepared review index to retrieve citable context from.section_policy(string) — Passage ordering: 'evidence_first' (default) or 'original_order'.sections(array) — Restrict retrieval to these article sections.session_id(string) — Optional staged session to scope retrieval to.table_mode(string) — Table rendering: 'off', 'preview' (default), or 'full'.verbosity(string) — Field verbosity: 'lean', 'standard' (default), or 'full'.
Database & release provenance
- Data mode
restored-database- Attestation
- attested-reference
- Data release tag
observed-2026.07.13- Release version
v7.1.10
sha256:e3a508657abbea8cb5a2282459f9313ff4f52aa27506f24d554292f9d2635295ghcr.io/berntpopp/pubtator-link@sha256:b2ab9657e9a7a3e4818ad58b3f183a205a5aba2c8f2117e964f4228a431064caReview the response
Keep the source record link and submitted identifiers with your notes. Record the dataset version and retrieval date when available. This page does not contain a captured PubTator3 response.