Developers

Public API

Everything published on EDA is open, and that includes the metadata. Read-only, no key needed for reasonable use, CC0 for the metadata itself.

The JSON API is planned, not live. The shape below is the design we are building toward and may change before launch. The OAI-PMH endpoint already responds.

Base URL

https://api.edapress.org

Endpoints

GET/v1/papers
List published papers, newest first.
subject, q, from, until, sort (recent | reads | citations), page, per_page (max 100)
GET/v1/papers/{id}
A single paper with authors, versions and identifiers.
GET/v1/papers/{id}/versions
Full version history for a paper.
GET/v1/papers/{id}/reviews
Published reviewer reports and decision letter, where the authors opted in.
GET/v1/subjects
Subject taxonomy with paper counts.
GET/v1/doi/{doi}
Resolve a DOI to its EDA record.

Limits

60 requests per minute without a key, 600 with a free key.
Responses are cached for five minutes; do not poll faster than that.
Bulk harvesting belongs on OAI-PMH, not the JSON API.
Papers still in review are never returned.

Example response

GET /v1/papers/eda-2026-0142

{
  "id": "eda-2026-0142",
  "title": "Attention Sparsity Patterns in Small Transformers Trained on Low-Resource Languages",
  "subject": "cs",
  "status": "published",
  "published_at": "2026-06-24",
  "version": 2,
  "license": "CC-BY-4.0",
  "doi": "10.0000/eda.2026.0142",
  "ark": "ark:/00000/eda-2026-0142",
  "authors": [
    { "name": "A. Mekonnen", "orcid": "0000-0002-1825-0097", "order": 1 },
    { "name": "J. Park", "orcid": null, "order": 2 }
  ],
  "links": {
    "landing": "https://edapress.org/paper/eda-2026-0142",
    "pdf": "https://cdn.edapress.org/papers/eda-2026-0142/v2.pdf"
  }
}

OAI-PMH

For harvesters. Supports Identify, ListMetadataFormats, ListSets, ListIdentifiers, ListRecords and GetRecord with Dublin Core. Sets map to subject codes.