FHIR IG Viewer
A standalone viewer for the Magentus Practice Management FHIR Implementation Guide. Built with Next.js, it mirrors the published IG and adds features that aren't possible within the standard FHIR IG publisher.
Why Build a Separate Viewer?
The official IG lives at fhir.dev.geniesolutions.io and is generated by the standard FHIR IG publisher toolchain. That pipeline is tightly coupled to the IG release process — any change to the viewer risks breaking the official deployment.
Building a separate viewer solves three problems:
Independent deployment. The viewer mirrors the real IG but ships on its own schedule. Updates to navigation, layout, or new features never touch the official publishing pipeline. The IG team can release profiles without worrying about viewer changes, and I can ship viewer improvements without waiting for an IG release.
AI features. The viewer includes Andy Clippy — a retro 90s-style AI assistant powered by Groq. It helps partners navigate FHIR profiles, explains concepts like Must Support and slicing, and links directly to relevant artifacts. This isn't possible in the standard IG publisher.
Freedom to experiment. Custom domain groupings, workflow diagrams, cross-references between profiles, and richer navigation. The standard publisher has a fixed output format. A custom viewer lets the information architecture match how people actually use the IG — browsing by clinical domain rather than alphabetical resource lists.
What I Built
- 68 profile detail pages with element trees, differentials, and raw JSON
- Domain landing pages grouping profiles by clinical workflow (bookings, diagnostics, referrals, core, admin, extensions)
- ValueSet and CodeSystem detail pages with concept tables
- Workflow diagrams for the bookings domain showing interaction and state models
- Andy Clippy: an AI chat assistant with full artifact context, streaming responses, and a Windows 98 retro UI
- All data sourced directly from the FHIR package JSON — no manual content entry