cdisc-library-mcp-server

cdisc-library-mcp-server

Provides a compact, task-oriented wrapper around the CDISC Library API for discovering, searching, and fetching Biomedical Concepts and SDTM Dataset Specializations.

Category
Visit Server

README

CDISC Library MCP Server (DXT)

This MCP server provides a compact, task‑oriented wrapper around the CDISC Library. It covers discovery (suggest + search), Biomedical Concepts (latest + versioned packages), and SDTM Dataset Specializations (latest + packages). Use it to search, fetch a concept or specialization with one call, then follow _links for deeper context. For reproducible results tied to an effective date, use package routes; for exploration, use latest routes.

About CDISC and the CDISC Library

  • CDISC: A global, non‑profit that develops data standards to streamline clinical research and enable interoperability. See the CDISC website.
  • CDISC Library: Programmatic access to CDISC standards and artifacts (e.g., SDTM, CDASH, ADaM, Controlled Terminology), including Biomedical Concepts and SDTM Dataset Specializations. Browse via the Library Browser and obtain API access from the Developer Portal.

Requirements

  • Node.js: >= 18
  • CDISC Library API key: subscription key from the Developer Portal

Get an API key

  1. Create/sign in to a CDISC Library account.
  2. Open the API Management Developer Portal and sign in with the same credentials.
  3. Retrieve your subscription key (API key) on the portal.
  4. Use the key as api-key in a header (recommended) or query string.

Run locally

export CDISC_API_KEY=your_key_here
npm start

Shell examples

  • bash/zsh:
export CDISC_API_KEY=your_key
export CDISC_AUTH_LOCATION=header   # or: query
npm start
  • fish:
set -x CDISC_API_KEY your_key
set -x CDISC_AUTH_LOCATION header   # or: query
npm start
  • PowerShell:
$env:CDISC_API_KEY = "your_key"
$env:CDISC_AUTH_LOCATION = "header"   # or: "query"
npm start

Configuration (environment variables)

  • CDISC_API_KEY (required): your subscription key
  • CDISC_AUTH_LOCATION (optional): header (default) or query
  • CDISC_AUTH_HEADER (optional): header name when using header auth, default api-key
  • CDISC_AUTH_QUERY (optional): query param name when using query auth, default api-key
  • CDISC_API_BASE_URL (optional): override server base URL, default https://api.library.cdisc.org
  • CDISC_OPENAPI_DIR (optional): directory for OpenAPI YAML; defaults to server/openapi. The manifest.json sets this to ${DXT_DIR}/server/openapi when packaged as DXT so the extension is self‑contained.

Caching and resilience

  • CDISC_CACHE_ENABLED (optional): set to 0 to disable caching. Default: enabled.
  • CDISC_CACHE_TTL_MS (optional): TTL for GET responses in milliseconds. Default: 60000.
  • CDISC_CACHE_MAX_ENTRIES (optional): max cache entries (LRU). Default: 500.
  • CDISC_RETRY_COUNT (optional): number of retries for transient 5xx/network errors. Default: 2.
  • CDISC_RETRY_BACKOFF_MS (optional): base backoff per attempt in ms. Default: 300.
  • CDISC_CACHE_DEBUG (optional): set to 1 to log cache/retry activity to stderr.
  • CDISC_CACHE_PERSIST_PATH (optional): file path to persist cache to disk (JSON). Default: disabled.
  • CDISC_CACHE_EXCLUDE_REGEX (optional): regex to exclude cache matches; defaults also exclude search/suggest.
  • CDISC_CACHE_INCLUDE_REGEX (optional): regex to force-include cache matches.

Package as a DXT extension

# Option A: one‑off
npm install -g @anthropic-ai/dxt
dxt pack        # creates cdisc-library-mcp.dxt in the project directory

# Option B: via npm script
npm run pack:dxt

Install in clients

  • Claude Desktop: open the generated .dxt file to install the local MCP server.
  • Any DXT‑compatible app can load this extension per its instructions. See the DXT repo: https://github.com/anthropics/dxt.

Tools exposed by the server

  • Generic:
    • list_operations(filter?)
    • call_operation(operationId, pathParams?, query?, body?, headers?, timeoutMs?)
  • Cache management:
    • cache.clear() — clear all in-memory cache entries
    • cache.invalidate(contains?, regex?) — invalidate entries by substring or regex match
  • Discovery:
    • search.suggest(q, top?, select?)
    • search.query(q, start?, pageSize?, facets?, filters?)
  • Biomedical Concepts (latest):
    • bc.list(category?), bc.get(conceptId), bc.categories()
  • Biomedical Concepts (packages):
    • bc.packages.list(), bc.packages.listConcepts(package), bc.packages.getConcept(package, conceptId)
  • SDTM Specializations (latest):
    • sdtm.list(domain?), sdtm.get(datasetSpecializationId), sdtm.domains(), sdtm.byBiomedicalConcept(conceptId)
  • SDTM Specializations (packages):
    • sdtm.packages.list(), sdtm.packages.listSpecializations(package), sdtm.packages.getSpecialization(package, datasetSpecialization)

Quick API reference

  • Auth & bases:

    • API key (header or query): api-key: <YOUR_KEY>
    • Library base: https://api.library.cdisc.org/api
    • Cosmos BC base: https://api.library.cdisc.org/api/cosmos/v2/mdr/bc
    • Cosmos SDTM Specializations base: https://api.library.cdisc.org/api/cosmos/v2/mdr/specializations
  • Discovery:

    • GET /mdr/suggest — params: q, top, select
    • GET /mdr/search — params: q (req), start, pageSize, facets, filters domain|codelist|conceptId|product|version|dataset|variable|standard
  • Biomedical Concepts:

    • Latest: /biomedicalconcepts, /biomedicalconcepts/{biomedicalconcept}, /categories
    • Version‑locked: /packages, /packages/{package}/biomedicalconcepts, /packages/{package}/biomedicalconcepts/{biomedicalconcept}
    • Key fields: _links, conceptId, href, categories, shortName, synonyms, resultScales, definition, coding, dataElementConcepts, ncitCode
  • SDTM Dataset Specializations:

    • Latest: /sdtm/datasetspecializations?domain=VS|LB|…, /sdtm/datasetspecializations/{dataset_specialization_id}, /sdtm/domains, by BC: /datasetspecializations?biomedicalconcept={conceptId}
    • Version‑locked: /sdtm/packages, /sdtm/packages/{package}/datasetspecializations, /sdtm/packages/{package}/datasetspecializations/{datasetspecialization}
    • Key fields: _links, datasetSpecializationId, domain, shortName, source, sdtmigStartVersion, sdtmigEndVersion, variables[]
    • variables[] keys: name, dataElementConceptId, isNonStandard, codelist, subsetCodelist, valueList, assignedTerm, role, dataType, length, format, significantDigits, relationship{subject,linkingPhrase,predicateTerm,object}, mandatoryVariable, mandatoryValue, originType, originSource, comparator, vlmTarget
  • Behavior:

    • Pagination on /mdr/search: start, pageSize
    • Errors surfaced as‑is: 401, 422, 404/406/5xx
    • Links: follow _links.self, _links.parentPackage, _links.parentBiomedicalConcept

Test the server

Scripts in package.json:

  • npm run test:mcp — smoke tests via a minimal MCP stdio client (scripts/test_mcp.js). Requires CDISC_API_KEY.
  • npm run test:plan — broader happy‑path + error scenarios (scripts/test_plan.js). Requires CDISC_API_KEY.

Troubleshooting

  • 401 Unauthorized: ensure CDISC_API_KEY is set; switch CDISC_AUTH_LOCATION between header and query if needed.
  • Timeouts/network errors: check connectivity to https://api.library.cdisc.org; set timeoutMs when using call_operation.
  • API changes: update the YAML specs under server/openapi, then restart. You can override the base via CDISC_API_BASE_URL.
  • Corporate proxies: configure your environment (e.g., HTTPS_PROXY, HTTP_PROXY).

Governance

  • License: MIT (see LICENSE.md).
  • Code of Conduct: see CODE_OF_CONDUCT.md.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured