Learning in public — this reference is being written in the open. Unfinished pages are excluded from search engines.
paged.IDML Reference

Paged

Paged is the open-source system that reads, renders, and edits IDML — a WebGPU render engine, a browser editor, an SDK, a paged.* scripting layer, a plugin platform, and a REST backend, all tracked by a live capability registry.

Tier: BeginnerBeginnerIexplanation

Paged is the open-source system that reads, renders, and edits IDML. This is the platform half of the documentation: the render engine and its WebGPU SDK, the paged.* scripting layer, the plugin SDK, the REST backend, and the live capability state. The other half is the IDML Reference — the format these tools act on.

In short the two references cross-link throughout. Every IDML construct carries a live badge showing how far Paged takes it — parsed, rendered, mutable, scriptable — sourced from the same capability registry surfaced here. Every API on this side points back to the format concept it reads or writes. Both are generated from the system's own artifacts, so they stay true as the code moves.

No demo session loaded.
A live, scrubbable recording of the real editor building a gradient — captured by running the journey tests, not hand-recorded.

Developer reference

  • SDKthe WebGPU rendering API (@paged-media/idml-viewer): load a document, render its pages, drive a pan/zoom camera. Read-only.
  • Scriptingthe paged.* host surface: read, mutate, and author documents from JavaScript, with a live playground.
  • Plugin SDKthe contract for extending the editor with tools, panels, commands, and content types.
  • REST APIthe editor backend's HTTP surface, generated from its OpenAPI spec.

The system

  • Pluginsthe five first-party content plugins: draw, web, image, sheets, data.
  • Repositorieshow the engine, editor, server, viewer, and state repos fit together.
  • Capability statethe live matrix of what works, at which layer, backed by test evidence. The system's source of truth for status.

The platform documentation is generated from the system itself wherever possible — the capability registry, the engine's scripting catalog, the plugin manifest schema, the server's OpenAPI spec, and the repos' commit history. When the software changes, these pages change on the next build. See how the capability state works.

On this page