Documentation

The Cascade Protocol provides a modular ontology framework for representing personal health and wellness data using Linked Data standards. This documentation covers the semantic vocabularies and SDK.

Getting Started

Get up and running with the Cascade Protocol in your language of choice.

TypeScript SDK

Full-featured SDK for Node.js and browser environments.

Read the guide

Python SDK Coming Soon

Python bindings for data science and research workflows.

CLI Quick Start Coming Soon

Command-line tools for working with Cascade pods and data.

Tools & SDKs

Install the tools you need to build with the Cascade Protocol.

CLI Reference Coming Soon

Command-line interface for managing pods, validating data, and working with Cascade vocabularies.

npm install -g @cascade-protocol/cli

TypeScript SDK

Full SDK for building Cascade Protocol applications in TypeScript and JavaScript.

npm install @cascade-protocol/sdk
View SDK docs

Swift SDK Early Access

Native Swift SDK for iOS, watchOS, and macOS applications. Currently in active development.

Contact for early access

Security & Compliance

The Cascade Protocol is designed with security and regulatory compliance at its core. Learn about encryption, access control, agent consent, and HIPAA-ready architecture.

Security & Compliance Guide

Comprehensive guide covering encryption at rest, consent-gated access, agent authorization, audit logging, and compliance controls.

Read the security guide

Specification

Formal specifications governing the Cascade Protocol data format, versioning, and pod structure.

Versioning Policy

How Cascade Protocol vocabularies are versioned, with semantic versioning rules and backward compatibility guarantees.

View versioning policy

Serialization Spec Coming Soon

Formal specification for RDF/Turtle serialization format, required prefixes, and canonical ordering.

Pod Structure Coming Soon

Directory layout, file naming conventions, and storage organization for Cascade data pods.

Agent Integration

Build AI agents that can securely access and process Cascade Protocol health data with explicit user consent.

MCP Server Setup

Run a local Model Context Protocol server for agent access to health data pods.

Setup guide

Agent Access Patterns

Consent-gated queries, provenance-tracked responses, and zero data egress architecture.

Learn more

Security for Agents

Audit trails, scope limitations, and compliance controls for AI agent interactions.

Security guide

Three-Layer Ontology Architecture

The Cascade Protocol organizes health data into three layers, each serving a distinct purpose:

1

Established Standards

Canonical clinical codes from FHIR, SNOMED CT, and LOINC. These are the universal language of healthcare systems.

2

Cascade Domain Vocabulary

Domain-specific properties (health:, clinical:, pots:, coverage:) that link to Layer 1 codes and add context for wellness devices, EHR data, screening protocols, and insurance/benefits.

3

Patient-Facing Summary

Patient-facing properties (checkup: for summaries, intake forms, visit preparation) — aggregated summaries, trend insights, and visit-prep documents designed for people, not systems.

Semantic Vocabularies

The Cascade Protocol uses modular RDF vocabularies built on W3C standards. Each vocabulary can be versioned independently while maintaining interoperability.

Core Vocabulary v2.3 Stable

Namespace: https://ns.cascadeprotocol.org/core/v1#
Prefix: cascade:

Cross-application vocabulary for schema versioning, data provenance, user identity, and patient demographics (PatientProfile, Address, PharmacyInfo, AdvanceDirectives). Used by all Cascade Protocol applications.

View Documentation Download TTL

Clinical Vocabulary v1.6 Stable

Namespace: https://ns.cascadeprotocol.org/clinical/v1#
Prefix: clinical:

Vocabulary for clinical documents and structured health records imported from EHR systems via Apple HealthKit. Supports medications, allergies, lab results, conditions, immunizations, procedures, vital signs, medication use episodes, supplements, insurance coverage records, and longitudinal lab tracking with FHIR alignment.

View Documentation Download TTL

Health Vocabulary v2.2 Stable

Namespace: https://ns.cascadeprotocol.org/health/v1#
Prefix: health:

Consumer wellness and device-generated health data — vital signs, activity, sleep, with SNOMED CT and LOINC standard code mappings for all metrics.

View Documentation Download TTL

POTS Vocabulary v1.3 Stable

Namespace: https://ns.cascadeprotocol.org/pots/v1#
Prefix: pots:

Domain-specific vocabulary for POTS home screening checks using the NASA Lean Test protocol. Supports age-adjusted thresholds, symptom logging, and blood pressure analysis.

View Documentation Download TTL

Coverage Vocabulary v1.0 Stable

Namespace: https://ns.cascadeprotocol.org/coverage/v1#
Prefix: coverage:

Layer 2 domain vocabulary for insurance, benefits, and financial health data. Provides standardized properties shared between patient-reported and EHR-imported insurance records, with FHIR Coverage alignment.

View Documentation Download TTL

Checkup Vocabulary v3.1 Stable

Namespace: https://ns.cascadeprotocol.org/checkup/v1#
Prefix: checkup:

Vocabulary for patient intake forms and pre-visit health data aggregation. Combines EHR-imported data, HealthKit wellness data, supplements, and manually entered information into comprehensive patient profiles.

View Documentation Download TTL

Diabetes Vocabulary Draft

Namespace: https://ns.cascadeprotocol.org/diabetes/v1#
Prefix: diabetes:

Unified vocabulary for diabetes and pre-diabetes management. Aggregates data from CGMs, glucose meters, insulin pumps, and lifestyle tracking devices into a coherent format with Time-in-Range metrics and pattern detection.

View Documentation Download TTL

ECG Vocabulary Coming Soon

Namespace: https://ns.cascadeprotocol.org/ecg/v1#
Prefix: ecg:

Electrocardiogram data from Apple Watch and medical devices. Planned for v2.0.

Ontology Explorer

Visualize the relationships between all Cascade Protocol vocabularies interactively. Explore the three-layer architecture, class hierarchies, and cross-vocabulary links as graphs and sortable tables.

Open Ontology Explorer

Interactive Guide

Explore the Cascade Protocol documentation interactively. Ask questions, get explanations, and learn about the vocabularies through conversation.

Powered by Google NotebookLM. Requires a Google account to access.

Open in Google NotebookLM

Quick Start

Here's a minimal example showing how the vocabularies work together:

@prefix cascade: <https://ns.cascadeprotocol.org/core/v1#> .
@prefix pots: <https://ns.cascadeprotocol.org/pots/v1#> .
@prefix prov: <http://www.w3.org/ns/prov#> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

<#pots-check-abc123> a pots:POTSCheckResult ;
    # Core vocabulary properties
    cascade:schemaVersion "1.2" ;
    cascade:dataProvenance cascade:ConsumerGenerated ;

    # POTS-specific properties
    pots:date "2025-12-05T14:30:00Z"^^xsd:dateTime ;
    pots:protocol "nasaLean" ;
    pots:potsThresholdMet "true"^^xsd:boolean ;

    # Provenance tracking
    prov:wasAttributedTo <https://id.cascadeprotocol.org/users/abc123> .

SDKs

TypeScript SDK Available

Full-featured SDK for building Cascade Protocol applications in TypeScript and JavaScript. Includes pod management, RDF serialization, encryption, and consent controls. 

npm install @cascade-protocol/sdk
View SDK docs

Swift SDK In Development

Native Swift SDK for iOS, watchOS, and macOS applications. Provides local encrypted Pod storage with AES-256-GCM, automatic RDF/Turtle serialization, HealthKit integration, WebID-based identity, and audit logging.

Contact us for early access.

Python SDK Planned

Python bindings for data science, research workflows, and server-side applications. Planned for future release.

Guiding Principles

  1. Established vocabularies first. Reference FHIR, LOINC, SNOMED CT, PROV-O, and other established standards before creating Cascade-specific terms. We only invent vocabulary when no standard exists.
  2. Three-layer ontology architecture. Every data type follows the Layer 1 (standards) → Layer 2 (domain) → Layer 3 (patient) pattern described above.
  3. SNOMED CT and LOINC mappings required. All clinical and wellness metrics must map to standard codes unless no code exists — in which case the gap is documented explicitly.
  4. Clear vocabulary separation. clinical: for EHR-derived facts, health: for consumer device data, pots: for POTS screening protocol (Layer 2), coverage: for insurance/benefits (Layer 2), checkup: for patient-facing summaries, core: for cross-cutting primitives, and domain prefixes for specialized protocols.
  5. Privacy-first, local-first. Data stored in Solid Pods with AES-256-GCM encryption at rest. No external API calls for core functionality.
  6. Provenance tracking. W3C PROV-O metadata on all data — distinguishing consumer-generated (device/user) from clinical-generated (EHR/provider) sources.

Ontology Objectives

  1. Interoperability: Any domain property should be traceable to established standard codes (SNOMED CT, LOINC, FHIR) so that Cascade data can round-trip with clinical systems.
  2. Completeness: Every data type serialized to Turtle has a corresponding vocabulary definition. No undefined terms.
  3. Consistency: All vocabularies follow the same structural patterns — OWL ontology headers, version info, changelogs, and SHACL shapes where appropriate.
  4. Discoverability: Every vocabulary has both a machine-readable .ttl file and a human-readable documentation page, linked from this index.
  5. Accuracy: Version numbers in documentation match the actual .ttl file versions. Changelogs reflect all changes.

Resources