Heratio Help Center article. Category: Technical / Integration.
Open-Data Catalogue (DCAT)
Overview
The platform publishes a single machine-readable catalogue of everything it offers as open data, using DCAT (the W3C Data Catalog Vocabulary, aligned with DCAT-AP). A data portal, harvester, or AI agent can fetch this one document and learn every dataset on offer, where to download each one, and in which formats - with no prior knowledge of the platform.
This catalogue is open data: no API key, read-only, published records only, and cross-origin (CORS) open so a browser app on any site can fetch it.
The catalogue endpoint
GET /data/catalog
The response is content-negotiated by the HTTP Accept header:
| Accept header | You get |
|---|---|
application/ld+json |
JSON-LD (the machine default) |
text/turtle |
Turtle |
application/rdf+xml |
RDF/XML |
text/html (a browser) |
a human-readable catalogue page |
You can also force a format with an explicit address:
GET /data/catalog.jsonld- JSON-LDGET /data/catalog.ttl- TurtleGET /data/catalog.rdf- RDF/XML
Or with a query string: ?format=jsonld | turtle | rdf | html. A plain command-line curl (which sends a catch-all Accept) receives JSON-LD.
What the catalogue describes
- A Catalog (
dcat:Catalog): the offering itself, with a title, description, licence (CC-BY-4.0), publisher, last-modified date, and a landing page. - One Dataset (
dcat:Dataset) per open-data surface: the bulk linked-data and CSV dumps, the per-record linked-data identity endpoints, the VoID dataset description, the OAI-PMH harvesting endpoint, the XML sitemaps, the syndication feeds, and the OpenAPI specification. Each dataset has a title, description, licence, publisher, and landing page. - One or more Distributions (
dcat:Distribution) per dataset: each concrete way to access the data, with the access URL (dcat:accessURL), a download URL where the data is a true file download (dcat:downloadURL), and the media type (dcat:mediaType). A dataset that serves several formats lists several distributions.
For surfaces that use a URL template (for example the per-record identity endpoint /id/{slug}), the distribution's access URL points at the protocol document and the template form is given in the distribution's description, so you still learn how to build a request.
How it relates to the protocol document
The catalogue and the protocol "capabilities" document (/open-data/protocol) describe the SAME set of open surfaces - one in DCAT, the other in a bespoke shape. They are generated from a single shared list, so they can never disagree. Use the DCAT catalogue when your tooling already speaks DCAT (data portals, CKAN, the European Data Portal); use the protocol document when you want one compact, self-explaining index. The protocol document links to the catalogue as its DCAT entry point.
Notes
- Every URL in the catalogue is built from the platform's own base address, so the document is correct on any deployment without configuration.
- The catalogue performs no database lookups and never fails over data: even an empty platform returns a valid, well-formed catalogue.
- The licence for the whole offering is Creative Commons Attribution 4.0 (CC-BY-4.0).