canonical: https://jentic.com/apis/core.ac.uk/core-ac-uk

# CORE API v2

Jentic publishes the only available OpenAPI specification for CORE API v2, keeping it validated and agent-ready. CORE is the world's largest aggregator of open access research, harvesting articles, journals and repositories and exposing them through a JSON API organised around three resource types. The 18 endpoints cover article retrieval by CORE ID with full-text PDF download, batch get and search, journal lookup by ISSN, repository metadata and similarity-based article discovery. Search supports an Elasticsearch-style query syntax over fields like title, authors, fullText and year, and authentication is an API key passed as a query parameter or header.

## For AI agents

Search and retrieve open access research articles, journals and repository metadata across millions of harvested papers, with full-text PDF access where available.

## Scope

Does not handle paywalled content licensing, citation export, or peer review workflows - use for open access article and repository discovery only.

## Capabilities

- Search articles using Elasticsearch-style queries across title, authors, full text and year
- Retrieve a specific article by CORE ID and download its full-text PDF when available
- Find near-duplicate articles for plagiarism detection or deduplication of search results
- Look up journals by ISSN to inspect their indexed coverage in CORE
- Browse repository metadata for institutional and subject repositories
- Find articles similar to a piece of input text for related-paper discovery

## Use cases

### Open Access Literature Search for Research Tools

Reference managers and research assistants can let users search CORE's open access corpus directly and retrieve metadata plus full text. The search endpoints accept Elasticsearch-style queries (e.g. title:psychology AND year:>2022) and the get endpoints pull full article records and PDFs for citation and reading.

Example prompt: GET /articles/search/{query} with the query 'title:transformer AND fullText:attention AND year:>2020' and return the top ten article CORE IDs with titles.

### Plagiarism and Similarity Detection

Editorial workflows and academic integrity tools can use CORE's similarity endpoint to compare submitted text against the open access corpus. POST /articles/similar takes input text and returns the most similar CORE articles, while POST /articles/dedup surfaces near-duplicate clusters useful for editorial review.

Example prompt: POST /articles/similar with the abstract of a submitted manuscript and return the five highest-scoring matches with their CORE IDs and titles.

### Repository and Journal Coverage Auditing

Library and scholarly communication teams can audit which institutional repositories and journals are indexed in CORE by listing repositories and journal records. POST /journals/get accepts batches of ISSNs to confirm coverage, supporting open access policy reporting and discoverability checks.

Example prompt: POST /journals/get with a list of department ISSNs and report which are indexed in CORE plus how many articles each has.

### AI Research Assistant with CORE Retrieval

An AI research agent can ground answers in open access literature by searching CORE through Jentic, fetching full text, and citing CORE IDs in its responses. Jentic isolates the API key, while the agent simply describes its query intent and consumes the JSON results into its retrieval-augmented pipeline.

Example prompt: Use Jentic to search CORE for 'climate adaptation small island states' filtered to year:>2023 and pull full text via /articles/get/{coreId}/download/pdf for the top three results.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| GET | /articles/get/{coreId} | Get an article by CORE ID |
| GET | /articles/get/{coreId}/download/pdf | Download full-text PDF by CORE ID |
| GET | /articles/search/{query} | Search articles with an Elasticsearch-style query |
| POST | /articles/search | Batch article search |
| POST | /articles/similar | Find articles similar to input text |
| POST | /articles/dedup | Get near-duplicate articles |
| GET | /journals/get/{issn} | Find a journal by ISSN |

## Key resources

- **Articles** — Search, batch get, single get, history, full-text PDF, similarity and deduplication endpoints
- **Journals** — Lookup by ISSN with single and batch operations
- **Repositories** — Repository metadata for institutional and subject repositories
- **All** — Global cross-resource search and retrieval methods

## Why Jentic

- **Setup:** Wiring the CORE API by hand means passing your apiKey as a query parameter and coding its article, journal, and repository search calls yourself. Through Jentic you install once, import the CORE API from the API Directory, store the key once, and your agent calls it.
- **Permission scoping:** CORE is a read-only research API where the article or query target travels in the request, so scope this by operations: allow the agent the reads it needs, such as searching articles and finding similar papers, and nothing beyond that. Each operation you credit the agent with stays inside that allowed set.
- **Credential handling:** Your CORE API key is stored once, encrypted, by your own Jentic One instance and injected into the query string at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'search CORE open access papers' or 'find similar CORE articles', and Jentic returns the matching articles, journals, or repositories operation with its query schema so the agent calls the right endpoint without browsing the reference docs.

## Related APIs

- **OpenAlex API** — Open scholarly knowledge graph with works, authors, institutions and concepts
- **Crossref API** — DOI-keyed citation metadata for the broader scholarly record
- **DataCite REST API** — DOI metadata for research datasets, complementing CORE's article focus

## FAQ

### Why is there no official OpenAPI spec for CORE API v2?

CORE does not publish an OpenAPI specification at this URL. Jentic generates and maintains this spec so that AI agents and developers can call CORE API v2 via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer.

### What authentication does the CORE API v2 use?

CORE uses an API key, accepted either as a query parameter or in a request header. Through Jentic the key sits in the credential vault and is attached to outbound requests at execution time.

### Can I download the full-text PDF of an article through the CORE API v2?

Yes. GET /articles/get/{coreId}/download/pdf returns the full-text PDF when CORE has it harvested. Not every article has a PDF available, so check the response status before parsing the body.

### How does CORE search query syntax work?

Search supports Elasticsearch-style queries over fields like title, description, fullText, authors, publisher, doi, oai, identifiers, language.name and year. Examples: title:psychology AND language.name:English, or repositoryDocument.metadataUpdated:>2024-01-01.

### What rate limits apply to the CORE API v2?

The OpenAPI spec does not declare formal rate limits. Use the batch endpoints POST /articles/get and POST /journals/get for bulk lookups instead of looping single GETs, and back off on HTTP 429.

### How do I find similar articles to a piece of text through Jentic with the CORE API v2?

Run pip install jentic, search Jentic with 'find similar CORE articles', load the operation schema for POST /articles/similar and execute it with the input text to receive ranked similar articles.

### Can I limit what my agent is allowed to do with the CORE API?

Yes. Because you run Jentic One yourself, your own rules decide which CORE operations and credentials the agent may use. CORE is a read-only research API, so you scope it by operation: grant only the reads the agent needs, such as searching articles with GET /articles/search/{query} or finding similar papers with POST /articles/similar, and withhold everything else like journal or repository lookups. Each operation you credit to the agent stays inside that allowed set, and your stored API key is attached at execution time rather than exposed to the agent.