phi-reader

offline viewer for healthcare C-CDA / CCD documents. makes no network requests of any kind.

Drop an .xml, .ccd, .cda or .zip onto it and it renders the chart — medications, allergies, labs, visit history — as something a person can actually read.

the network stack is not involved

This is the design constraint everything else follows from. phi-reader makes no network requests of any kind: no telemetry, no cloud processing, no update check, no remotely-loaded fonts or scripts. Clinical documents are opened, parsed and rendered on the machine in front of you.

• Parsing runs in the main process; rendering happens in a hardened, sandboxed renderer.
• Sanitisation is vendored locally rather than pulled from a CDN, so there's no third party in the path even for a library file.

a document viewer that phones home is a document viewer that exfiltrates, however good the intentions

how it renders

Two paths, chosen per document:

• When the archive carries an HL7 stylesheet, the XSLT transform is applied — the presentation the document's author intended.
• When it doesn't, a fallback narrative parser walks <structuredBody> directly and builds the chart from the structured entries.

The second path is what makes it useful in practice. Plenty of exported records arrive without a stylesheet and render as raw XML everywhere else.

specs

Platform	Windows, macOS, Linux
Formats	C-CDA, CCD — .xml, .ccd, .cda, .zip
Built with	Electron
Network access	none
Status	private beta