Four ways in, depending on what you are building. Every path runs entirely on your machine: no account, no API key, and no network call is required to read or write a Pod.
TypeScript SDK
Read and write Pods from Node or the browser. Typed models for every clinical and wellness class, SHACL validation, and JSON-LD context handling.
Start with TypeScript npmPython SDK
The same model surface for data and research work, with dataframe-friendly readers over a Pod's Turtle files.
Start with Python PyPICommand Line
Create a Pod, import FHIR data, validate against SHACL shapes, query records, and run an MCP server, without writing any code.
Start with the CLI npmCascade Agent
Give an AI assistant read and write access to a Pod over MCP. Agent-written records carry AIGenerated provenance automatically, and every operation is logged.
Start with the agent Agent examplesBefore you build
The vocabularies define what a record can say, and the specification defines how it is written to disk. Both are worth a skim before you model your own data.
Vocabularies
Every OWL class and property, with SHACL shapes and JSON-LD contexts.
Browse vocabulariesSpecification
Serialization rules, Pod directory structure, and the versioning policy.
Read the specification