About OpenRx
OpenRx exists because the United States already publishes nearly everything you could want to know about a medication — and publishes it in fourteen incompatible formats, across three agencies, keyed on five different identifiers. This site joins those sources into one schema and one interface.
What OpenRx is
An index, faithfully assembled
Every fact on every page is reproduced from a U.S. government dataset and carries a source marker you can follow. Clinical text is quoted verbatim from the Structured Product Labeling the manufacturer filed with the FDA. Nothing is paraphrased, interpreted, summarised by a language model, or invented.
A join across agencies
The regulatory spine is FDA's NDC Directory. Onto it are joined Drugs@FDA for approval history, the Orange Book for therapeutic equivalence, patents and exclusivity, the Purple Book for biologics, Animal Drugs @ FDA for veterinary products, openFDA for labelling, recalls and adverse events, and the NLM's DailyMed mappings for clinical vocabulary.
What OpenRx is not
Not medical advice. OpenRx is not a clinical decision support tool and must not be used to make decisions about care. Dosing, interactions, contraindications and suitability depend on the individual patient and on clinical judgement.
Not the official labelling. The prescribing information shown here is a reproduction of what was on file with the FDA at the time this build was generated. Labels change. The manufacturer's current package insert is the authoritative document.
Not a safety ranking. Adverse event report counts are raw counts. They do not measure risk, incidence or causality.
See the full disclaimer and the methodology.
How the site is built
One schema
Twenty-seven tables in SQLite, applied unchanged to Cloudflare D1. Provenance is a first-class table: every dataset carries its version, release date, byte count and SHA-256.
Prerendered static
Astro emits every page at build time and Cloudflare Workers Static Assets serves them from the edge — no origin, no server render, no cold start on a page view.
Offline search
Pagefind indexes the whole site into static fragments that the browser downloads and queries locally. Search works without round-trips and without a search vendor.