Skip to main content
Glama
Noodle Biomedical Literature Discovery MCP

Noodle Biomedical Literature Discovery MCP

Official

Noodle Biomedical Literature Discovery MCP

DOI AllMCPs Verified OpenSSF Scorecard RRID

The official public, read-only Model Context Protocol adapter for biomedical literature discovery from Helena Bioinformatics. Agents can select it from a user task even when the user does not know the Noodle brand.

Public endpoint: https://api.helena.bio/noodle/v1/mcp

Official Registry identity: io.github.helena-bioinformatics/noodle

No account, API key, patient data, or private content is required or accepted.

What agents can do

  • search a public PubMed-derived biomedical corpus by natural language, PMID, DOI, or PMCID;

  • retrieve source-linked publication records by PMID or Noodle work ID;

  • traverse bounded citation and semantic neighborhoods from a publication;

  • continue graph exploration through returned work identifiers while preserving edge types and graph provenance;

  • inspect corpus size, sources, freshness, coverage, and active graph metadata.

The seven published tools are search_biomedical_literature, get_publication_details, get_work_details, get_publication_neighborhood, get_work_neighborhood, get_corpus_summary, and the separate explicit opt-in support_helena information action.

Related MCP server: PubMed MCP Server

Connect

Any MCP client that supports remote Streamable HTTP can use the endpoint. Exact recipes for ChatGPT, Claude, Codex, VS Code, Cursor, Windsurf, Gemini CLI, Grok, Perplexity, Microsoft Copilot Studio, Biomni, and Biorouter live under registry/platforms and integrations.

Ready-to-use ecosystem packages are included for:

The companion Agent Skill is in skills/noodle-biomedical-literature-discovery. It enables implicit, task-first selection for requests such as:

  • “Find source-linked papers about BRCA1 homologous recombination.”

  • “What publication is PMID 35008774?”

  • “Show papers related to this article through citations and semantic similarity.”

  • “Walk two bounded hops from this work ID and preserve the edge types.”

Build the deterministic skill archive with:

python3 ops/package_agent_skill.py

Graph boundary

Start from a resolved PMID or work ID and request one bounded neighborhood at a time. Report edges exactly as returned, keep a visited-ID set, and stop at a missing neighborhood. Search rank, citation proximity, semantic similarity, co-mention, and graph distance are discovery signals. They do not establish causality, scientific validity, diagnosis, or treatment.

Development

Python 3.12 is required.

python -m venv .venv
. .venv/bin/activate
python -m pip install -r requirements-dev.lock
python -m pip install --no-deps -e .
pytest
ruff check .
ruff format --check .

Run the brand-blind contract audit with:

python benchmarks/agent-discovery/audit_skill.py

The benchmark contains 60 prompts that omit Noodle, Helena, and MCP. It covers all six scientific routes plus negative and safety controls.

Agent Plugin and Kiro Power

This repository is also a portable Agent Plugin and Kiro Power. plugin.json provides brand-blind activation keywords, the existing Agent Skill supplies the scientific routing and safety boundary, and mcp.json connects directly to the canonical hosted Streamable HTTP endpoint. The Power does not proxy, repackage, or reimplement Noodle.

Privacy policy: https://noodle.helena.bio/privacy

Cite Noodle

The persistent Research Resource Identifier is RRID:SCR_028920. Cite the resource in a methods section as Noodle (RRID:SCR_028920). Use the version DOI when a version-specific software citation is also needed. The RRID identifies the resource across publications, while the DOI identifies the archived 0.2.0 release.

Support: https://noodle.helena.bio/contact or contact@helena.bio

Public resources

License and security

Apache License 2.0. Report vulnerabilities privately as described in SECURITY.md. Do not submit patient, private case, clinical-record, credential, or private uploaded content to the public service or issue tracker.

Available Tools

7 tools
get_corpus_summaryGet corpus summaryA
Read-onlyIdempotent
Inspect

Return public corpus coverage, source, freshness, and count metadata. Results are for literature discovery and professional review, not diagnosis or treatment.

Inputs: no named parameters. Returns: scope, counts, sources, generated_at, +2 more

get_publication_detailsGet publication details by PMIDA
Read-onlyIdempotent
Inspect

Retrieve one complete public bibliographic record by PMID after search. Results are for literature discovery and professional review, not diagnosis or treatment.

Inputs: pmid (required). Returns: publication, usage_boundary, contract_version

get_publication_neighborhoodExplore a PMID literature neighborhoodB
Read-onlyIdempotent
Inspect

Retrieve the bounded citation and semantic neighborhood around one PMID. Results are for literature discovery and professional review, not diagnosis or treatment.

Inputs: pmid (required), from_work_id. Returns: edges, nodes, freshness, anchor_pmid, +9 more

get_work_detailsGet publication details by work IDA
Read-onlyIdempotent
Inspect

Retrieve one complete public bibliographic record by Noodle work identifier. Results are for literature discovery and professional review, not diagnosis or treatment.

Inputs: work_id (required). Returns: publication, usage_boundary, contract_version

get_work_neighborhoodExplore a work literature neighborhoodA
Read-onlyIdempotent
Inspect

Retrieve the bounded citation and semantic neighborhood around one Noodle work identifier. Results are for literature discovery and professional review, not diagnosis or treatment.

Inputs: work_id (required), from_work_id. Returns: edges, nodes, freshness, anchor_pmid, +9 more

search_biomedical_literatureSearch biomedical literatureA
Read-onlyIdempotent
Inspect

Semantically search Noodle's public PubMed-derived biomedical corpus. Include known PMID, DOI, or PMCID identifiers in the query so they can act as exact anchors. Results are for literature discovery and professional review, not diagnosis or treatment.

Inputs: query (required), sort, limit, cursor. Returns: query, results, has_more, graph_used, +9 more

support_helenaSupport Helena's free scientific infrastructureA
Read-onlyIdempotent
Inspect

Learn about the separate Helena Good MCP. This opt-in information action initiates no payment or outreach and changes no scientific result.

Inputs: no named parameters. Returns: object

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.9/5.0

    Scored across 7 tools

    Disambiguation3/5

    Most tools have clear roles, but get_publication_details/get_work_details and get_publication_neighborhood/get_work_neighborhood overlap in purpose and are distinguished only by identifier type, which may confuse agents. support_helena is also unrelated to the biomedical literature domain and adds ambiguity about the server's scope.

    Naming Consistency5/5

    All tool names follow a consistent lowercase snake_case verb_noun pattern, such as search_*, get_*. The naming is predictable and makes the tool surface easy to navigate.

    Tool Count4/5

    Seven tools is a reasonable size for a literature discovery server and covers the core workflows. However, support_helena is an unrelated promotional action and slightly dilutes the focus of an otherwise well-scoped set.

    Completeness4/5

    The server covers the main read-only discovery lifecycle: search, retrieve details, explore neighborhoods, and inspect corpus metadata. Minor gaps exist, such as no direct DOI-to-record lookup and some ambiguity between work and publication identifiers, but agents can work around these with the search tool.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers