OpenAlex, the free and fully open (CC0-licensed) scholarly metadata catalog maintained by the nonprofit OurResearch, changed the terms of access to its hosted REST API in early 2026: every request now requires a registered API key, and most operations are billed on a usage basis rather than offered unmetered. The change affects how the API is called, not the underlying data itself — the full dataset remains freely downloadable as a CC0 snapshot from a public AWS S3 bucket, with no account or key required for bulk access.
What changed
Per OurResearch’s own announcement on the OpenAlex blog, the hosted API now requires an API key for essentially all requests: "you’ll need an API key for all requests." A small number of calls remain possible without a key for demonstration purposes, but OurResearch is explicit that this unauthenticated path is not suitable for production use. This is a shift from the API’s prior posture, where a key was recommended (and needed to raise rate limits) but not strictly required to make basic calls.
Alongside the key requirement, OpenAlex introduced usage-based pricing for the API itself. Every API key receives a daily free-usage allowance, described by OurResearch as roughly $1 of free usage per day. Within that model, different call types are priced differently:
- Single-record lookups by ID or DOI — free, with no volume cap.
- List and filter calls — a small per-call cost (on the order of $0.0001 per call, i.e. about $0.10 per 1,000 calls).
- Search calls — a higher per-call cost (on the order of $0.001 per call, i.e. about $1.00 per 1,000 calls).
- PDF/XML full-text downloads — the highest per-call cost among documented operations (on the order of $0.01 per call, i.e. about $10.00 per 1,000 calls).
Users who need volume beyond the daily free allowance — or who want data sync, curation dashboards, or other organizational features — can move to one of OurResearch’s paid annual plans (marketed as Member, Member+, and Supporter tiers), arranged directly with OurResearch rather than self-serve. Institutions with heavier needs, such as monthly full snapshots or daily change files delivered outside the S3 bucket, are directed to contact OurResearch’s sales channel.
Why this matters for research administration
OpenAlex has become a default, no-cost substitute or supplement to proprietary bibliographic databases for many research offices, CRIS integrations, and bibliometrics workflows precisely because it combined broad Microsoft Academic Graph-derived coverage with an open, unmetered API. That calculus now has a wrinkle: applications, dashboards, or CRIS/RIMS integrations built against OpenAlex’s REST endpoints for live queries — as opposed to a periodic bulk-data refresh — may accrue real, if generally modest, per-call costs once free-tier usage patterns exceed the daily allowance. Search and full-text download operations are priced noticeably higher than simple ID/DOI lookups and paginated list/filter calls, so the cost impact will vary a lot by use case: an integration doing occasional entity lookups is unlikely to notice, while one running frequent full-text-search queries across a large corpus is more likely to approach or exceed the free allowance.
For institutions and vendors with OpenAlex-dependent tooling, the practical near-term steps are the same as with any metadata-provider pricing change: register an API key if one isn’t already in use (this is now required regardless of usage volume), review which of your integration’s calls fall into the priced categories above, and confirm whether current usage is likely to stay inside the daily free allowance or warrant a paid plan.
What OurResearch says the change is meant to do
OurResearch has framed the shift as sustaining a "free data, paid service" model: the underlying scholarly metadata stays CC0 and freely downloadable in bulk, while the convenience of a hosted, queryable API — infrastructure that OurResearch operates and pays to run — is what carries the new usage-based cost. That framing is consistent with the Principles of Open Scholarly Infrastructure (POSI), which OpenAlex has previously cited as a commitment, and which do not require that every mode of access to open infrastructure be free of charge — only that the underlying data and its terms remain open.
What remains unconfirmed
OurResearch’s announcement does not spell out a formal grace period or migration deadline for API consumers who were previously calling the API without a key, nor does it give an exact clock-time cutover date beyond referencing an earlier announcement. Organizations running production integrations should treat their own API dashboards and the current OpenAlex help documentation as the authoritative, up-to-date source for exact current rates and any transition allowances, since usage-based pricing on hosted infrastructure is the kind of policy that gets refined after initial rollout.
Related CASRAI coverage
See CASRAI’s guides to OpenAlex: What It Is and How It Works for background on the catalog’s scope, entities, and licensing, and OpenAlex API: Endpoints, Authentication, and Pagination for the mechanics of authenticated calls, pagination, and rate limits, both updated to reflect the current key requirement. For how OpenAlex compares to other bibliographic sources, see Scopus vs Web of Science vs OpenAlex and Metadata Search Engines for Scholarly Research.
