Europe PMC
europepmc.org · Debrid · 1 record · JSON
Europe PMC Debrid
caveatanonymousno cardDrequest-rate
| Free tier | REST access to 'over 33 million publications', including '10.2 million full text articles and 6.5 million open access articles', with JSON or XML responses and pageSize up to 1000. unit unlimited; metered on none; pageSize=1000 returned 786,907 bytes of JSON in one response here; pageSize=100 returned 88,669 bytes. |
|---|---|
| Price | USD 0 |
| Per-object cap | not applicable: one JSON/XML response per request, paged. The largest single response measured was 786,907 bytes at pageSize=1000. |
| Retention | the index is continuously updated; the bulk FTP tree carries dated snapshots (weekly for metadata, per-release for the OA subset) |
| Egress | Free and unmetered from the EBI endpoint. No key, no account, no registration mentioned anywhere on the web-service page. free_allowance: no byte allowance published |
| Identity | anonymous No key and no account: an anonymous GET returned 200 and 1,408 bytes. Nothing on the web-service page asks for registration. |
| Limits |
The bulk path is FTP and it did not answer from this network: https://europepmc.org/ftp/oa/ returned 403 and ftp://europepmc.org/oa/ timed out after 20 s. That is a reachability fact about for a user, not a statement that the mirror is gone. |
| Binding | request-rate With the FTP mirror unreachable here, the API is the only path, and it is one page per call with no published ceiling: at 787 KB per pageSize=1000 call, a TiB would take about 1.4 million calls, so the call rate decides the day's bytes, not any byte quota. |
| Interface | rest, ftp, cli GET https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=...&format=json&pageSize=1000&cursorMark=...; eight bulk content types are listed at /downloads and served over FTP. |
| Tooling | curl/python-requests for the API; lftp or curl for the FTP tree. No SDK needed. |
| Composes with | Use Europe PMC as the index (PMIDs, PMCIDs, licence flags, full-text availability) and pmc-oa-opendata as the byte store; the API's pageSize=1000 gives about 787 KB of metadata per call without an account. |
| Reliability | No throttling signal was seen at the low rate used here; the FTP tree is the fragile part - one 403 on HTTPS and a timeout on port 21 make the bulk mirror look absent when it may only be filtered. |
| Notes | Caveat is the bulk mirror, not the API. Metadata is effectively unlimited through pageSize=1000; full text at TiB scale depends on the FTP tree, which this network could not open. |
| Evidence rung | documented The API was exercised at two page sizes with byte counts, and the bulk-path failures were observed. The 33M/10.2M/6.5M counts are the vendor's own page, not independently measured. |
| Checked | 2026-10-24 (live); stale on 2026-11-23 |
| Parsed numbers |
|
Evidence (4)
| date | source | what it proved |
|---|---|---|
| 2026-10-23 | europepmc.org/RestfulWebService | 'over 33 million publications', '10.2 million full text articles and 6.5 million open access articles', JSON/XML output |
| 2026-10-23 | europepmc.org/downloads | eight bulk content types (OA subset, author manuscripts, preprints, metadata, ID mappings, text-mined terms) served over FTP |
| 2026-10-23 | probe:https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=OPEN_ACCESS:Y&format=json&pageSize=1000 | 200, 786,907 bytes in one anonymous response - pageSize=1000 is accepted and returns real records |
| 2026-10-23 | probe:https://europepmc.org/ftp/oa/ | 403 over HTTPS - the FTP tree has no HTTPS mirror at that path |
More fields (5)
{
"page_1000_bytes": 786907,
"page_100_bytes": 88669,
"publications": 33000000,
"oa_fulltext": 6500000,
"ftp_https_status": 403
}