Self-archiving is the practice of depositing a copy of your own research output into an open, freely accessible repository, separately from wherever the publisher makes the formal published version available. It is the mechanism behind green open access: rather than paying a publisher to make an article open at the point of publication (gold OA), the author (or a delegate, such as a library or research office) makes a permitted version openly available themselves. Done correctly, self-archiving costs nothing beyond staff time and satisfies most funder and institutional open-access mandates without requiring an article-processing charge.
This guide walks through the three practical checks every self-archiving deposit requires — which manuscript version you are allowed to post, whether an embargo period applies, and where to put it — plus how the result maps onto common funder mandates.
Step 1: Identify which manuscript version you’re allowed to deposit
Publishers rarely allow self-archiving of the final typeset, published article. Almost all publisher self-archiving policies specify a particular manuscript version, and depositing the wrong one is the most common compliance mistake. NISO’s Journal Article Versions (JAV) framework (NISO RP-8-2008) is the standard reference vocabulary publishers use to describe these stages, and the three that matter most for self-archiving are:
- Author’s Original (AO) — the manuscript as submitted, before peer review. Sometimes called a preprint.
- Accepted Manuscript (AM) — the version that incorporates changes made during peer review and has been accepted for publication, but has not yet been copyedited or typeset by the publisher. This is the version most green OA policies (including most funder mandates) require or permit for self-archiving.
- Version of Record (VoR) — the publisher’s final, copyedited and typeset version, usually carrying the journal’s own formatting, pagination, and DOI. Self-archiving the VoR is permitted far less often, and typically only under a specific license (see Rights Retention Strategy, below) or after a longer embargo.
Before depositing anything, confirm which of these your specific publisher agreement permits — this is usually stated in the copyright transfer or license-to-publish agreement you signed, or in the publisher’s standard self-archiving policy for that journal.
Step 2: Check whether an embargo period applies
Many publishers permit self-archiving of the Accepted Manuscript only after an embargo period has elapsed — commonly somewhere between 6 and 24 months after publication, though the specific length varies by publisher and, in some cases, by discipline. Two practical ways to check:
- Open Policy Finder (formerly Sherpa Romeo) — a searchable database, run by Jisc, that aggregates publisher and journal self-archiving policies, including which version is permitted and any embargo length. It remains the standard first stop for checking a journal’s green OA terms, even though the tool itself has been rebranded from its long-standing “Sherpa Romeo” name.
- The publisher’s own self-archiving or “author rights” policy page — the authoritative source when it and an aggregator disagree, since aggregator entries can lag a publisher’s actual current policy.
Some funder mandates now override publisher embargoes entirely. Since July 2025, the NIH Public Access Policy requires NIH-funded Accepted Manuscripts to be deposited in PubMed Central with zero embargo, made publicly available at the article’s official publication date, regardless of what a publisher’s own self-archiving policy would otherwise permit. The NSF has adopted a comparable zero-embargo requirement for NSF-funded Accepted Manuscripts deposited in NSF-PAR, for awards made on or after January 22, 2026. Where a funder mandate and a publisher’s embargo conflict, check your funder’s specific public-access policy — many funder agreements now take precedence, but the mechanics (and any negotiated exceptions) vary by publisher and funder.
Step 3: Choose where to deposit
Once you know which version you’re allowed to post and any embargo has passed (or been waived by a funder mandate), you have three realistic deposit options:
- Institutional repository (IR) — the repository run by your university or research institution. This is the deposit location most funder and institutional OA policies expect by default, and the one most consistently indexed by discovery tools and most durably maintained over the long term.
- Subject (disciplinary) repository — a repository organized around a field rather than an institution, such as arXiv (physics, mathematics, computer science), bioRxiv or medRxiv (life sciences and health sciences), or PubMed Central for biomedical and life-sciences literature funded by NIH and other participating funders. Subject repositories are a good fit when your discipline has a well-established one and your funder or publisher policy accepts it as a valid deposit location.
- A personal or departmental website — permitted under some publisher policies, but the weakest option for actually satisfying a funder mandate or supporting long-term discoverability: personal pages are not consistently indexed, are not preserved when a researcher changes institutions, and most funder policies do not recognize them as compliant deposit locations on their own. Posting to a commercial academic social network (ResearchGate, Academia.edu) is generally not a substitute for a genuine institutional or subject repository deposit either — these platforms are not indexed the same way, and many publisher agreements draw an explicit distinction between “public repositories” and general-purpose social platforms.
You can look up which repositories a publisher or funder recognizes via directories such as OpenDOAR (a registry of open-access repositories) alongside the version/embargo information in Open Policy Finder.
How self-archiving fulfills funder green OA mandates
Most research funder open-access policies are satisfied by a correctly executed self-archiving deposit — that is the entire point of a “green” route as distinct from a “gold” (pay-to-publish-open) one. In practice, meeting a funder’s green OA mandate means all three steps above have to align with that specific funder’s requirements simultaneously: the right manuscript version, deposited within (or without) the embargo window the funder allows, in a repository the funder recognizes. A deposit that satisfies a publisher’s own self-archiving policy does not automatically satisfy a funder mandate layered on top of it — the funder’s requirements are usually the stricter of the two and should be checked independently.
A related but distinct mechanism worth understanding alongside self-archiving is the Rights Retention Strategy (RRS), associated with cOAlition S and Plan S. Rather than relying on whatever self-archiving terms a publisher’s own policy happens to grant, RRS has the author apply a CC BY (or equivalent) license to their Accepted Manuscript at the point of submission — via a standard funder-provided rights-retention statement in the manuscript or cover letter — so the right to immediately self-archive under an open license exists independently of the publisher’s own embargo or version policy. Where it applies, RRS effectively removes the embargo question from Step 2 for the Accepted Manuscript.
Common mistakes to avoid
- Depositing the Version of Record when only the Accepted Manuscript is permitted — this is a copyright violation in most publisher agreements, not just a technicality.
- Ignoring the embargo shown in Open Policy Finder or the publisher’s policy and depositing immediately, when a delay is required.
- Depositing only to a personal website or academic social network and assuming that satisfies a funder mandate — check the funder’s policy for which repository types it actually recognizes.
- Assuming one deposit satisfies every obligation — a single Accepted Manuscript upload to an institutional repository may need to be supplemented with a separate deposit to a funder-mandated repository (e.g., PubMed Central for NIH-funded work) if the institutional repository alone isn’t the funder’s recognized compliance route.
Frequently asked questions
What version of my paper can I self-archive?
It depends on the publisher’s specific policy for that journal, checked via Open Policy Finder or the publisher’s own author-rights page. Most policies permit the Accepted Manuscript (the peer-reviewed, pre-copyedit version); fewer permit the Author’s Original (pre-review preprint) without restriction, and fewer still permit the Version of Record without a specific license such as Rights Retention.
Do I need the publisher’s permission to self-archive?
You need to follow whatever self-archiving terms are in the copyright transfer or license-to-publish agreement you signed, or a rights-retention statement applied before submission if your funder requires one. Most established publishers have a standing self-archiving policy that doesn’t require a separate, individual permission request — but the terms of that standing policy (version, embargo, repository type) are binding.
How long is a typical green OA embargo?
It varies by publisher and sometimes by discipline, commonly in the range of 6 to 24 months, though an increasing number of funder-driven mandates (such as the current NIH and NSF Accepted Manuscript requirements) now require zero embargo regardless of what the publisher’s own default policy states. Always check the specific journal and the specific funder mandate rather than assuming a typical figure applies.
Does self-archiving on ResearchGate or Academia.edu count?
Generally no. Most funder and institutional OA policies specifically require deposit in a recognized institutional or subject repository, not a commercial academic social network, and many publisher self-archiving policies draw the same distinction explicitly.
Where can I find my publisher’s self-archiving policy?
Open Policy Finder (formerly Sherpa Romeo) aggregates policies across most major publishers and is the standard starting point; the publisher’s own author-rights or self-archiving policy page for that specific journal is the authoritative source if the two differ.
