Rxnorm MCP Server

RxNorm — Standardized Drug Concepts

The National Library of Medicine's standardized drug nomenclature. Every prescription, OTC drug, and ingredient in the US healthcare system has an RxCUI (a stable concept ID). RxNorm normalizes drug naming chaos: "Tylenol" and "acetaminophen" and "APAP" all resolve to the same ingredient concept. Free, no auth.

Part of Pipeworx — an MCP gateway connecting AI agents to 1476+ live data sources.

Why this matters for AI agents

Drug data is notoriously messy. Brand names, generic names, salt forms, dose strengths, dosage forms — same molecule, dozens of strings. RxNorm gives you canonical IDs you can use to cross-reference FDA data, drug interaction databases, EHR systems, and clinical trials.

Three core flows:

1. Resolve a name. "What's Ozempic?" → rxnorm_search({name: "ozempic"}) → concept groups by term type (SBD, BN, IN, etc.) with RxCUIs.

2. Get drug properties. "What's this RxCUI?" → rxnorm_get_properties({rxcui}) → ingredient name, strength, dose form, brand status.

3. Drug interactions — not available. NLM discontinued the RxNav interaction API in January 2024 and nothing replaced it, so this pack has no rxnorm_interactions tool and Pipeworx has no drug-drug interaction source. pharma_safety_report returns the FDA label's own Drug Interactions section, which documents one drug rather than checking a pair; a drug missing from that text has not been cleared.

For full drug-safety synthesis combining RxNorm + FDA + ClinicalTrials, use [compare_entities({type: "drug", values})] or the pharma_drug_profile compound.

Term types (TTYs) — important to understand

RxNorm has multiple concepts per drug, each at different abstraction levels:

For cross-source linking, always prefer the IN (ingredient) RxCUI when one exists. Brand drugs in different countries have different SBDs; the IN is global. resolve_entity({type: "drug"}) picks the canonical RxCUI in priority order IN > PIN > BN > SBD > first available.

Citable URIs

pipeworx://rxnorm/concept/{rxcui}
pipeworx://rxnorm/concept/{rxcui}/interactions

RxCUIs are stable and shareable.

Update cadence

RxNorm publishes monthly updates. New drugs appear within a month of FDA approval; deprecated products are flagged but kept in the database for historical lookup. Pipeworx caches with a 24-hour TTL.

Common pitfalls

• Search returns SBD-only for brand queries. Searching "Ozempic" returns Semantic Branded Drug concepts, not the ingredient. To get to "semaglutide," either search the generic name directly or call rxnorm_related({rxcui, tty: "IN"}) on a brand RxCUI. The latter currently has bugs in some configurations — see error recovery.
• Spelling matters more than you'd think. "metformin hydrochloride" and "metformin HCl" return slightly different concept sets. The search is forgiving for common abbreviations but exact for less common ones.
• NDCs are different. National Drug Codes (NDCs) are FDA-issued package-level identifiers, not RxCUIs. Use rxnorm_ndc({rxcui}) to bridge if you need NDC for prescription-fill data.
• Combination products. A drug like "Kombiglyze" contains metformin + saxagliptin. Searching "metformin" may surface combination products as the top hit, which isn't what you want for ingredient-level analysis. Filter for SCD or IN, not just take the first result.
• Discontinued drugs. Concepts persist after a drug is withdrawn. The obsolete flag in the full record tells you whether the drug is currently available.

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "rxnorm": {
      "url": "https://gateway.pipeworx.io/rxnorm/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/rxnorm/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Both URLs reach the same gateway and the same 1476+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:

ask_pipeworx({ question: "your question about Rxnorm data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides
• pipeworx.io

License

MIT