For machines

Everything here is open access under CC BY 4.0, served as static files with open CORS, and free to read, harvest and reuse with attribution. There is no API key, no rate limit and nothing to register for.

Full text

Articles from 2026 onward are published as validated structured HTML, so the text you retrieve is the text of record rather than an extraction from a PDF. Each one is also rendered to Markdown, which is usually what you want:

https://dcpapers-data.dublincore.org/articles/{issue}/{id}/files/dcmi-{id}.md

It is generated from the same source and the same numbering as the HTML and the typeset PDF, so a figure or reference number means the same thing in all three. It opens with YAML front matter carrying the title, authors with ORCIDs, DOI, licence and a version_of_record flag.

Availability is declared per article as files.markdown, and advertised as a Signposting item. Older proceedings are PDF-only, so check the field rather than assuming the URL resolves.

Bulk and structured data

The whole corpus is static JSON. Prefer it to scraping pages.

manifest.json — collection statistics and publisher metadata
index.json — every article id mapped to its path
issues/index.json — every proceedings volume
llms.txt — orientation for language models and agents

Per article, alongside article.json: content/article.html for full text, content/refs.yml for the bibliography, and files/dcmi-{id}.pdf for the typeset PDF.

Harvesting and discovery

OAI-PMH — oai_dc metadata, one set per proceedings, 100 records per page.
ResourceSync — capability, resource and change lists for synchronisation.
Signposting — FAIR Level 2 relations (cite-as, item, author, license) in the link header, the page head and a JSON link set at /article/{id}/linkset.json.
Sitemap — every article and issue page.

Citing and identity

Resolve identity through the DOI, never the title — titles are not unique across years. DOIs take the form https://doi.org/10.23106/dcmi.{id}, where the id is the one in the article URL. Every article page offers BibTeX, RIS and CSL-JSON exports at /article/{id}/references.{ext}.

Check the publication stage before citing. An issue circulating for correction is an uncorrected proof, not the Version of Record, and its cite-as relation is deliberately withheld. Each article carries its stage at custom.publicationStage; each issue carries versionOfRecord.

The published record is authoritative. An article absent from index.json is not published, and should not be inferred or reconstructed from other sources.