EpiSAP Data Sources & Attribution Policy v1

Document: EpiSAP Data Sources & Attribution Policy v1.0
Effective date: 30 April 2026
Issuer: EPISAPIENT PTY LTD (ACN 664 593 542)
Status: Published transparency notice. Not a Gate-required acceptance document. Provided to fulfil the transparency commitments in the Privacy Policy and to evidence good-faith compliance with the licence terms of every external data source EpiSAP relies on.

1. Purpose

EpiSAP's Publication Watch service (also called "Pub-Alerts") monitors external public-record data sources for publications, preprints, and patent filings that may overlap with a User's locked Idea. This Policy explains: which sources we read from, the licence we rely on for each, what we store internally, what we display to Users, and how we attribute every source.

This Policy is informational. Nothing in it expands, narrows, or otherwise modifies the EpiSAP Platform Terms of Use or Privacy Policy.

2. Scope: what we read versus what we display

2.1 EpiSAP reads publicly-available metadata from the sources listed in section 3 in order to (a) identify candidate publications relevant to a locked Idea, and (b) run an internal AI-overlap analysis on those candidates.

2.2 EpiSAP stores internally, for each candidate that crosses our relevance threshold, the following fields: title, authors, DOI (where available), publication date, source name, canonical URL, and (where the source provides one) the abstract. The abstract is used solely as input to the AI-overlap analysis. Internal abstract length is capped at 5,000 characters per record and the field is not displayed to Users.

2.3 EpiSAP displays to the User, in the Pub-Alerts dashboard, only: the publication title, authors, DOI, publication date, source name, a canonical link out to the original source, our own AI-generated overlap analysis, and our internal severity rating. EpiSAP does not display, redistribute, or republish abstracts or full text from any external source.

2.4 EpiSAP does not mirror, cache, redistribute, or republish corpora from any external source. Each scan queries the live source through its public API; we keep only records that produce alerts, and only the metadata listed in clause 2.2.

3. Source inventory

Source What we read Licence / legal basis Cost
Crossref REST API
(operated by Crossref, US 501(c)(6) DOI registrar)
DOI metadata for journal articles and for bioRxiv / medRxiv preprints (filtered by type:posted-content, container-title:Biorxiv or Medrxiv). Crossref distributes its metadata under CC0 1.0 (public domain) for substantially all fields. Use of the REST API is governed by Crossref's Public Data Terms; we operate within the "polite pool" by sending a User-Agent header that includes our contact email. Free.
NCBI E-utilities
(operated by the US National Library of Medicine, part of NIH)
PubMed records: title, authors, abstract, journal, publication date, PMID, MeSH terms. PubMed metadata is a US Government work in the public domain. NCBI E-utilities are made available for programmatic access under the NCBI Disclaimer and Copyright Notice; rate limits (3 req/sec without an API key, 10 req/sec with) are observed. Free.
OpenAlex
(operated by OurResearch, US 501(c)(3))
Scholarly work metadata: titles, authors, abstracts (where present), institutions, publication years. OpenAlex is in the public domain, made available under CC0 1.0. Free.
arXiv API
(operated by Cornell University)
Preprint metadata: title, authors, abstract, categories, version dates. Programmatic access is permitted under the arXiv API Terms of Use for research, education, and indexing purposes. Rate limit (1 request / 3 seconds) is observed. Free.
bioRxiv / medRxiv
(operated by Cold Spring Harbor Laboratory, with Yale and BMJ for medRxiv)
Reached primarily through Crossref (above). If a direct fallback is ever required, content is licensed by the author at submission, with the majority under CC-BY 4.0 and the remainder under CC-BY-NC, CC-BY-ND, or CC0. Per-preprint Creative Commons licence chosen by the author; all permit indexing and informational citation with attribution. Free.

3.2 Additional public-record sources may be added in future releases — for example, the World Intellectual Property Organization's PATENTSCOPE, the United States Patent and Trademark Office (USPTO) PatFT/AppFT, the European Patent Office Espacenet OPS, and the Indian Patent Office (Office of the Controller General of Patents, Designs & Trade Marks; "CGPDTM"). When any such source is added, this Policy will be updated and the change will be reflected in the issued version of this Policy.

4. Attribution and link-out policy

4.1 Every Pub-Alerts entry shown to a User identifies the source by name (for example, "PUBMED", "BIORXIV / MEDRXIV", "ARXIV", "OPENALEX") and provides a clearly-labelled hyperlink to the canonical record at the source. Where a DOI is available, the canonical link is the DOI URL.

4.2 EpiSAP encourages Users to follow the canonical link to read the full publication on the source's site, where the source's own access controls, licence terms, and publisher rights apply.

5. Internal use and the no-mirroring rule

5.1 EpiSAP treats every external source as an authoritative, live resource. We do not maintain a local mirror, search index, or full-text cache of any external corpus.

5.2 The internal storage described in clause 2.2 is limited to (a) the records that produced an alert for a User, and (b) the metadata fields necessary to render the alert and preserve the evidentiary chain that connects the alert to the User's locked Idea. This is materially smaller than substantial extraction from any source database and is consistent with permitted indexing use under the licences listed in section 3.

5.3 We do not redistribute, republish, syndicate, or sell metadata or content sourced from any of the providers listed in section 3.

6. Trademarks

6.1 "Crossref", "PubMed", "MEDLINE", "OpenAlex", "arXiv", "bioRxiv", and "medRxiv" are the trademarks or service marks of their respective operators. EpiSAP uses these names only descriptively, to identify the source of metadata, and does not claim affiliation with, sponsorship by, or endorsement from any of these operators.

7. Take-down requests

7.1 If a publisher, author, or rights-holder believes that a specific Pub-Alerts entry exceeds the scope of permitted use described in this Policy, they may contact us at copyright@episapvault.com. We will review the request promptly, and where appropriate suppress the affected entry pending resolution.

7.2 The full notice-and-counter-notice procedure for copyright complaints is set out in the EpiSAP Repeat Infringer Policy.

8. Changes

We may update this Policy from time to time. The effective date and version number at the top of this Policy will be updated when changes are made. Material changes will be notified to Users via the same channels described in the Privacy Policy.

— End of EpiSAP Data Sources & Attribution Policy v1.0 —