A developer-first framework for structured, patient-owned health data.
Cascade Protocol provides semantic vocabularies, serialization formats, and developer tools for building health applications where patients own and control their data. All processing is local-first — no data leaves the user's machine.
| Repository | Description | Status |
|---|---|---|
| spec | Ontology files, SHACL shapes, serialization specification, JSON-LD context | Stable |
| conformance | Conformance test fixtures, reference Patient Pod | Stable |
| cascade-cli | CLI for validating, converting, querying, and managing health data Pods | v0.3.3 |
| sdk-typescript | TypeScript SDK (@the-cascade-protocol/sdk) |
v1.1.0 |
| sdk-python | Python SDK (cascade-protocol) |
v1.0.1 |
| cascade-agent | Natural language interface for the Cascade Protocol CLI | v0.1.7 |
# Install the CLI
npm install -g @the-cascade-protocol/cli
# Import FHIR data into a Pod
cascade pod init ./my-pod
cascade pod import ./my-pod patient.json
# Query by data type
cascade pod query ./my-pod --medications --conditions --lab-results
# Validate against the Cascade spec
cascade validate ./my-pod- Local-first: Zero network calls during operation. All data stays on your machine.
- Patient-owned: Data stored in portable Pods that users control.
- Provenance-tracked: Every record carries its origin — clinical, device, self-reported, or AI-generated.
- Semantic: Built on RDF/OWL with W3C PROV-O provenance. SPARQL-ready knowledge graphs.
- Open: Apache 2.0 (code) / CC-BY 4.0 (specifications). No proprietary dependencies.
- cascadeprotocol.org — Documentation & getting started guides
- CLI Guide
- Cascade Agent Guide
- Security & Compliance
Code repositories: Apache License 2.0 Specification & documentation: CC BY 4.0