Introducing healthcare.dev

Every team that builds healthcare software rebuilds the same plumbing. Someone downloads the ICD-10-CM release from CMS, writes a parser for its fixed-width files, loads it into a table, and promises to do it again next October. Someone else does the same for HCPCS every quarter, for the FDA's NDC Directory every day, and for RxNorm every month. Then the formats drift, the cron job breaks, and a claim fails validation because a code retired six months ago.

healthcare.dev does that work once, for everyone, and serves the result through one API.

What's in the first release

• Codes: ICD-10-CM, ICD-10-PCS and HCPCS Level II, with search, hierarchy, batch lookups and as_of dates, so a claim from 2019 validates against the codes of 2019.
• Drugs: the FDA NDC Directory and RxNorm, linked, so any NDC in any format resolves to its RxNorm concepts.
• Freshness you can see: every response carries the source, the release and the license; a public status endpoint shows how current each dataset is.
• An MCP server so agents in Claude, Cursor, VS Code and Codex can use the data directly, and docs with llms.txt and Markdown pages so coding agents can learn the API on their own.

What we won't do

We won't handle patient data. healthcare.dev is reference data only, and our terms forbid sending PHI. That keeps you and us out of HIPAA business associate scope, and it keeps the product simple enough to sign up for in two minutes without a sales call.

We also won't ship data we're not allowed to: no CPT, no CDT, and nothing licensed until the license is signed.

What's next

Provider data from NPPES is next, then FDA labels, recalls and shortages, LOINC, fee schedules and, later, hospital and payer price transparency data. Follow along in the changelog, or get a free key and make your first call.