How the Scite MCP Works With Publisher Content

No agreement, no full text. Unless a publisher has an explicit Scite MCP amendment, no full-text content beyond openly licensed material (CC-BY and equivalent permissive licenses) is ever shown — not to the user, and not to the AI tool they’re using. Subscription and paywalled text stays behind the paywall, full stop. Even open access articles under restrictive licenses (non-commercial, no-derivatives) are excluded from AI-readable text.

Every result carries a DOI. Each article surfaced to the user includes its DOI and a resolved access link, so the researcher is routed to the article on the publisher’s platform — the version of record — through their institution’s subscription or a legitimate per-article purchase. Scite sends readers to you; it doesn’t replace you.

Indexing drives discoverability. Fully indexed articles (metadata + abstracts + full-text citation snippets) consistently surface higher in AI-mediated results. Which means fully-indexed content will more likely appear in the user’s search results and lead to higher link-outs to the publisher’s version of record.

We provide publishers with usage and denial data to help drive subscription discussions. Scite provides publishers and institutional customers with the same organization-level and journal-title-level usage and denial reporting: which journals researchers accessed through AI tools, and which they tried to access but couldn’t. Both sides of the subscription conversation work from the same complete picture — turned-away demand becomes visible evidence for new subscriptions and per-article sales, not lost signal.

Everything is logged and read-only. Access is limited to authenticated Scite subscribers, individually credentialed, usage-capped, and fully auditable.