Introduction

Hypercube connects to your data and serves it as markdown pages an agent can read and navigate like a small website.

The model

• Resource — a data source. Either static (tables you define and fill by hand) or dynamic (a Postgres database Hypercube syncs: table schemas plus sample rows, cached in the application database).
• Table — a named collection of fields and records inside a resource.
• Cube — a set of markdown pages, one of them the entry page.
• Page — a JIM document: markdown in which {{ … }} holds a JavaScript expression evaluated against your resources, for example {{ resources["Shop"]["orders"].length }}.

The API

A cube is served over HTTP:

• GET /c/:cube — the cube's entry page
• GET /c/:cube/:page — a specific page

Responses are markdown. Links between pages resolve to full URLs, so an agent can follow them from the entry page through the whole cube.

Where to go

• Installation — run Hypercube from source or with Docker.
• Connecting resources — bring your data in.
• Building cubes — turn it into navigable pages.
• Writing pages — JIM, the page format.
• API — the endpoints a cube is served on.