Structured data

Translating content into a structured representation does more than power a better reading experience. Once an article is structured data rather than a flat document, the whole corpus becomes something you can connect, query, and build on in new ways.

An open internal representation¶

Curvenote works with and toward the Open Exchange Architecture (OXA), an open way to represent structured scientific content. Rather than treating an article as opaque text or a PDF, OXA represents its parts — sections, figures, equations, citations, authors, and identifiers — as connected, addressable structure.

That structure is the foundation for everything below: when each element of a document is explicitly represented and linked, the content can be reasoned about, not just rendered.

New possibilities¶

A knowledge and citation graph¶

Because citations, authors, funders, institutions, and resources are resolved to identifiers and enhanced, your corpus becomes a graph you can traverse — not a pile of documents. You can follow how works cite one another, find articles connected to a dataset or institution, and surface relationships across publications that are otherwise invisible.

Querying and APIs¶

Structured content can be queried programmatically. Instead of scraping documents, you can ask questions of the content and its metadata directly — powering feeds, dashboards, and integrations, or pulling exactly the slice of the corpus you need into another system.

AI and agent access¶

Structured, machine-readable content is well suited to retrieval and agentic use. OXA gives AI systems a faithful, connected view of your corpus to ground on — enabling discovery, summarization, and analysis over real content rather than guesswork. You stay in control of where this is allowed; see Bots & AI traffic.

Understand the content your system holds¶

The same structured representation underlies the Curvenote Scientific Content Management System (SCMS). If you operate an SCMS, this is not only about reading experiences for published content — it is a way to interrogate, search, and understand the content your system has in place: what you hold, how it connects, and where the gaps are. See the Curvenote SCMS documentation to learn more.