Cascade, on one page.

A person's health record is in twenty places, and none of them is theirs.

Portals, devices and apps each hold a piece. The law now gives people their records, but what arrives is a FHIR bundle or a C-CDA archive built for exchange between institutions. Each app that tries to bring the pieces together invents its own data model, so the record can't move on to the next app.

Cascade is a shared model for that record, kept on the person's side, written in the codes clinicians and their systems already use.

A folder of files, in three layers.

A pod is a folder of plain RDF (Turtle) files on the person's device, laid out to Solid conventions. Any tool can open it. The CLI can encrypt it at rest with a passphrase.

Nothing is silently overwritten.

What arrived from a source stays as it arrived. Corrections, merges of duplicates from two hospitals, and retractions are written as a separate layer on top. You can always see the original and what changed it.

Many collectors, one record.

Some of the most useful health data never reaches the chart: a teacher's rating of a child's attention, a standing heart rate at home, a symptom diary. Purpose-built apps collect it well. When they write Cascade, it can land in the same record as the clinical data, with its provenance, instead of in another silo.

That record is what lets families manage care, learn what their records say, and take a bigger part in decisions. See the apps doing this.