icdwise
Verified ICD-10-CM code lookup & validation for AI agents — official descriptions, not guesses.
README
<div align="center">
<img src="https://qinisolabs.github.io/icdwise/logo.svg" width="96" height="96" alt="Qiniso" />
icdwise
Verified ICD-10-CM code lookup & validation for AI agents — official descriptions, not guesses.
Verified, trustworthy data tools for AI agents. "Qiniso" means "truth" in Zulu.
Website · npm · MCP endpoint · MCP Registry
</div>
Ask an LLM what an ICD-10-CM code means and it will answer confidently — and often wrongly: the wrong laterality, the wrong severity, the wrong condition entirely, and it will happily invent a description for a code that doesn't exist. icdwise looks the code up in the official U.S. ICD-10-CM code set and returns the real description — or an honest "not found" instead of a guess.
Asked to describe real ICD-10-CM codes, a frontier model with no tools was wrong ~38% of the time — e.g. it called
H40.1131"severe stage, right eye" (it's mild, bilateral) andT63.011A"ingested mushrooms" (it's rattlesnake venom) — and it fabricated a description for a code that doesn't exist. icdwise: the official text, or "not found." Never a guess.
Add it to Claude
Settings → Connectors → Add custom connector, and paste — no login, no key:
https://icdwise.qinisolabs.workers.dev/mcp
Stateless, reads no user data, requires no secrets. Prefer to run it locally over stdio? Add { "command": "npx", "args": ["-y", "icdwise"] } under mcpServers in your client config.
Use it as a library
npm i icdwise
import { lookupIcd10, validateIcd10, searchIcd10 } from "icdwise";
lookupIcd10("E11.9").description; // "Type 2 diabetes mellitus without complications"
lookupIcd10("H40.1131").description; // "Primary open-angle glaucoma, bilateral, mild stage"
validateIcd10("E11.99").valid; // false — well-formed but not a real code
searchIcd10("generalized anxiety"); // → [{ code: "F41.1", description: "Generalized anxiety disorder" }, ...]
Codes are accepted with or without the dot (E11.9 or E119). A well-formed code that isn't in the official set returns found: false with a clear note — it never invents a description.
Tools — 3
| Tool | What it answers |
|---|---|
| lookup_icd10 | The official description of a code (+ canonical form and 3-char category) |
| validate_icd10 | Is this a real ICD-10-CM code? (well-formed and in the official set) |
| search_icd10 | Reverse lookup — find the code(s) for a condition by keywords |
Data
ICD-10-CM is U.S. public-domain data (NCHS/CMS). The full code set (~74,000 codes) is bundled and generated from the official CMS release via npm run build-data <icd10cm-codes-YYYY.txt> (see scripts/build-data.mjs); every response reports the datasetVersion it used. Descriptions are the official text — the curated, versioned dataset kept current is the moat.
What it is not
- Not medical advice, and not a determination of billability, coverage, or clinical appropriateness.
- Not ICD-10 (WHO) or ICD-11 — this is ICD-10-CM (the U.S. Clinical Modification).
- Not a guesser — unknown/retired/non-leaf codes return an honest "not found", never a fabricated description.
Architecture
A single TypeScript package exposing one MCP server over two transports — stdio (local / npx) and a Cloudflare Worker (hosted edge endpoint) — both driven by the same core.ts tool definitions, which also power the importable library.
npm install
npm run build
npm test
License
Apache-2.0. ICD-10-CM data is U.S. public domain (NCHS/CMS); see NOTICE.
Recommended Servers
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.
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.
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.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
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.
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.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.