How to get an arXiv endorsement in 2026, and where else to post

An arXiv endorsement is how arXiv decides that a new submitter belongs to the research community of a subject area, and you need one before your first paper on arXiv or in a new category. Since 21 January 2026, an institutional email address no longer qualifies you on its own. There are now two paths: automatic endorsement, which needs both an institutional email address and authorship of a paper already accepted in the same endorsement domain, or a personal endorsement from an established arXiv author in that domain. arXiv staff cannot endorse you or waive the requirement. If neither path is open to you, repositories and preprint servers such as Zenodo, OSF Preprints, Preprints.org and Research Square post research without an endorsement, each under its own rules.

This guide covers what changed, how the request works, who can endorse, how to ask, what happens after you submit and where else to post. Every policy below was checked on 15 October 2026 on the provider's own pages or, where a page blocked automated access, on its most recent Internet Archive copy, noted in the sources. These rules have changed several times in a year, so check the linked page before you rely on a detail.

What changed in arXiv's endorsement policy in 2026?

On 21 January 2026, arXiv stopped treating an institutional email address as enough for endorsement in every section. Its announcement says arXiv "will no longer accept institutional email addresses (i.e., an email address associated with an academic or research institution) as the sole qualifier of endorsement for new authors". The reason it gives is "stemming the flood of low-quality, non-scientific submissions to arxiv.org". Institutional email addresses, arXiv wrote, "are no longer a sufficient credential for determining minimum research competence or endorsement."

Mathematics went first. On 10 December 2025, automatic endorsement in all Math categories began to require an institutional email address and "previous authorship on an existing paper which has been accepted to the arXiv Mathematics section". Mathematical Physics (math-ph) belongs to Physics and followed Physics rules. Six weeks later the same model applied everywhere.

arXiv describes endorsement domains as "high-level subject areas or related categories on arXiv". In its words, "Most high-level subject areas (e.g., hep-th, cond-mat, q-bio) are currently endorsement domains, with the notable exception of physics, in which individual subject classes (e.g., phys.acc-phys, phys.med-ph) are endorsement domains."

Four more points:

• Existing endorsements stay. "Authors who were previously endorsed to submit to a particular category will still be endorsed in that category." The change mostly affects people submitting to a category for the first time.
• Staff cannot help you skip it. "arXiv staff cannot waive endorsement requirements or provide a personal endorsement for authors." arXiv points questions about the policy to its Endorsement Policy Feedback Form.
• Claim your papers. For the automatic path, the paper must be linked to your account. If a co-author submitted it, you can claim it at once with its paper password, or use the Request Ownership form, which "involves verification by our staff and may take a couple of days" (Authority Records).
• Endorsement can be taken away. "Endorsement can also be revoked if an author violates arXiv policies or the arXiv code of conduct."

How do you know if you need an arXiv endorsement?

You find out when you start a submission: if your account is not endorsed for the category you choose, arXiv tells you then and emails you instructions. Its support article puts it this way: "The process is initiated during the article submission process. At that time users are notified if endorsement is required and are sent an email with instructions." We found no separate endorsement status page in arXiv's help pages as of 15 October 2026; the submission itself is the check.

The request works like this, according to arXiv's endorsement page:

1. Start a new submission and select the category you want.
2. If you need an endorsement, arXiv sends an endorsement request email. It "will include a link you can provide to prospective endorsers".
3. The endorser receives a "six-character alphanumeric endorsement code" from you and enters it on arXiv's endorsement form, which needs an arXiv login.
4. The endorser either endorses you or explicitly declines, which "will be recorded as a negative vote of endorsement". Someone who does nothing has abstained.
5. "At least one positive endorsement is required per endorsement category to be considered endorsed for that category."

Whether a person endorsed you or not is "private between you and the arXiv administrators", though arXiv notes you might be able to infer it. There is no charge: arXiv's submission guidelines say "submission to arXiv is free for authors".

Who can endorse on arXiv?

An endorser is an arXiv author with enough recent papers in the endorsement domain, who is registered as the author of those papers and is endorsed in that area. arXiv's endorsement page says the number of papers "depends on the particular subject area", and that "we only count papers that have been submitted between three months and five years ago." Its support article on endorsers gives the range: "typically between 2 and 5 articles that have been originally made public in arXiv within the period of 5 years to 3 months ago."

Three further conditions apply:

• The papers must be linked to the endorser's account. "If you would like to be able to endorse, the most important thing that you can do is make sure that you are registered as an author of your papers."
• The endorser needs an endorsement in that area too: "you must also have an active positive endorsement to that area yourself before you may endorse for that subject classification."
• Status is automatic. "Endorser status is completely automated and not something that has to be requested, nor can arXiv staff make you an endorser at your request." Seniority, affiliation and papers published outside arXiv do not count towards it, and arXiv reserves the right to suspend a person's ability to endorse.

So a well-known professor who rarely posts to arXiv may not be able to endorse you, while a postdoc with several recent arXiv papers in your endorsement domain may.

How do you find an arXiv endorser?

Start with people who know you or your work: arXiv says it is best to find an endorser who "you know personally" and who "is knowledgeable in the subject area of your work". For graduate students, it suggests "your thesis advisor or another professor in your department/institution working in your field who is also an active author on arXiv." Co-authors, colleagues and people you have met at seminars or conferences are the next places to look.

If you know no one, arXiv describes how to search (endorsement page):

1. Look at recent papers in your reference list that are on arXiv. People whose work you build on are well placed to judge whether your paper fits the field.
2. Open each paper's abstract page. At the bottom is a link titled "Which authors of this paper are endorsers?" As of 15 October 2026, the link asks you to log in to arXiv before it shows the list.
3. arXiv's help page says you can find the email address "just under the Submission history heading". On the abstract page that is a "view email" link for the person who submitted the paper; for a co-author, use the address in the paper itself.
4. Write to eligible endorsers individually, not in bulk, and forward the endorsement request email. You may include "a link to your ORCID profile, other scholarly profile, or a copy of your intended submission if you are comfortable doing so."

What an endorser is asked to check. arXiv tells endorsers: "The endorsement process is not peer review. You should know the person that you endorse or you should see the paper that the person intends to submit. We do not expect you to read the paper in detail, or verify that the work is correct, but you should check that the paper is appropriate for the subject area." Endorsers should say no if "the author is unfamiliar with the basic facts of the field, or if the work is entirely disconnected with current work in the area." A paper they receive is to be treated "as privileged information", and arXiv asks endorsers who feel uncomfortable for any reason not to endorse but to ask the person to find someone else.

A short, respectful request. One that makes the check easy might contain:

1. A subject line naming the category, such as arXiv endorsement request, math.NT.
2. One or two sentences on who you are and why you are writing to this person, for example the paper of theirs that you cite.
3. The title, the abstract and the PDF, or a link to it, so they can see the paper fits the category.
4. The endorsement link or code from arXiv's email.
5. A line that says plainly that a no, or no reply, is fine.

What to avoid. arXiv is explicit: "it is inappropriate to email large numbers of potential endorsers at once, or to repeatedly email the same endorser with a request for endorsement." Do not press someone who has not replied, and do not ask anyone to endorse on another person's behalf: endorsers are told to endorse only "authors seeking to submit their own work, not to third-parties or proxies." A request to arXiv staff will not help either, since they cannot grant or waive endorsement.

Does an endorsement mean arXiv will accept your paper?

No. Endorsement lets you submit; moderation decides whether and where the paper appears. arXiv's moderation page says "arXiv reserves the right to reclassify or decline any submission", and also that "the arXiv moderation process is not a peer-review process." Moderators are "volunteer subject matter experts with terminal degrees in their field".

These are the rules most likely to matter after endorsement, as of 15 October 2026:

On AI, arXiv reminds authors that "by signing their name as an author of a paper, they each individually take full responsibility for all its contents, irrespective of how the contents were generated." Its October 2026 rate-limit announcement adds that arXiv policy permits authors to use AI as a tool in their research "as long as it is disclosed and the research meets arXiv's standards of scholarly interest and advances the field of research." Our guide to AI disclosure shows how to word such a statement.

The Computer Science rule is not technically new, arXiv says: review articles and position papers "are not (and have never been)" among its accepted content types and were accepted at moderator discretion. Each category's moderators may follow with a similar practice, and arXiv says it will announce any that do.

Can an independent researcher get an arXiv endorsement?

Yes, through personal endorsement, and arXiv does not require an institution. Its registration help says: "If you are unaffiliated you may enter 'Independent' as your organization." Without an institutional email address, though, the automatic path is closed, so you need an established author in your endorsement domain to endorse you. If you have co-authored papers already on arXiv, claim them anyway: arXiv recommends that you "claim ownership of all of them and keep your arXiv account as up-to-date as possible", and its December 2025 post for Mathematics told co-authors without an institutional email to claim their papers and, "If still unable to submit", to seek personal endorsement.

Do not borrow an affiliation to get through. arXiv's identity and affiliation policy says "Claimed affiliation should be current in the conventional sense: e.g., physical presence, funding, e-mail address, etc." and that "Misrepresentation of identity or affiliation, for any reason, is possible grounds for immediate and permanent suspension."

What should an independent researcher write as an affiliation?

Write that you are an independent researcher, where the journal allows it, and never an institution you no longer belong to. You can publish a research paper without an affiliation at publishers whose author policies provide for it, and at least two say so explicitly.

• IOP Publishing: "If a researcher is not affiliated to an institution or company, they can list themselves as 'independent researcher'." (Affiliations)
• Taylor & Francis: "If you do not have a current relevant institutional affiliation you should state your independent status." Note the exception on the same page: "For non-research articles, you must list your current institutional affiliation." Taylor & Francis lists "Affiliation misrepresentation" among examples of misconduct (Editorial Policies).

If you move to another institution before the article is published, Taylor & Francis asks for "the affiliation where the work was conducted" and a note with your current one. Rules differ between journals, so read each journal's instructions for authors before you submit. Some preprint servers ask for affiliations too, as the table below shows.

Where can you post a preprint without an arXiv endorsement?

General repositories and other preprint servers post research without an endorsement, though most screen submissions first. The table shows what each says on its own pages, as of 15 October 2026.

A few notes on the table.

OSF is a host, not one server. OSF Preprints is "a network of community-run preprint servers", including PsyArXiv, SocArXiv, MetaArxiv and Law Archive, and you choose the one that fits your field. OSF is phasing out its Projects between November 2026 and February 2027, but its transition guide says preprints and registrations are unaffected.

Field servers have their own scope. medRxiv is not intended for, among other things, narrative reviews, "opinion pieces and hypotheses that lack data". If your paper is a hypothesis, our guide on where to publish one compares the venues that accept it.

HAL's documentation is in French. Its about page says it serves researchers "affilié.es à des institutions académiques", public or private, including foreign ones, and "L'utilisation de HAL est gratuite" (use of HAL is free).

Not in the table. TechRxiv and ESS Open Archive blocked automated access on 15 October 2026 and we could not confirm their current terms on their own pages, so we left them out. ResearchGate is not in the table either: it describes itself as a network whose members "use ResearchGate to connect, collaborate, and share their work" (About).

Zenodo or arXiv: which should you use?

Use arXiv when you can get endorsed and want the paper moderated and placed in a subject category; use Zenodo when you need a DOI for any research output, with no endorsement or moderation. arXiv is a moderated preprint server: you need an endorsement, and volunteer moderators check each submission and may move it to "the most relevant category" or decline it. Zenodo is a general repository: "Anyone may register", it accepts "All types of research artifacts" at "Any status", and it registers a DOI for every published record. Zenodo provides all information "as-is", in its own words, and removes records outside its scope after they appear rather than screening them before.

You can use both: Zenodo for data, code and supplementary files, and arXiv or a field server for the paper. A DOI makes a deposit findable and citable. It says nothing about whether the work has been reviewed or is correct.

What to do if arXiv put your paper on hold or declined it

If arXiv put your paper on hold, wait and do not resubmit; if arXiv declined it, read the message, fix what can be fixed and appeal only with something new. A status of "on hold" means your submission is still being checked, not that it was refused. arXiv's status page says "If you notice your submission is on hold, no action is required on your part at this time", that the process "may take several days", and: "Please do not make a duplicate submission while your work is on hold."

If arXiv declined the paper, expect little explanation: "arXiv staff and moderators cannot give feedback on the submission." What helps:

1. Read the message you received. It may name a category, a content type or a format problem.
2. Check the paper against the policies above. It should be a complete research article in the right category, with an English version and with any AI use reported.
3. Fix what can be fixed, then ask a colleague in the field to read it before you try again.
4. Appeal only with something new. Appeals go through arXiv's user support portal, not to moderators directly, and must explain why the moderators should reconsider, with "a pdf copy of your work" for declined submissions (appeals). "Most appeals are resolved within a two week period", and a denied appeal is final.

Two cautions. A rejected submission still counts towards your two submissions that month. And "Submitters who have had works previously delayed or declined by arXiv should anticipate closer scrutiny on future submissions." Nothing guarantees that a revised paper will pass. Posting it on another server from the table above is a legitimate choice.

Where a Papex record fits

Papex is a record registry, not a journal or a preprint server. It does no peer review, needs no endorsement and mints no DOI. If you need a DOI, use one of the repositories above; a Papex record can sit alongside it.

Anyone with a Papex account can start a private draft, as a Hypothesis, Discovery or Publication record. The record stays a private draft until you sign it. The signature links you to that exact version, and only a salted fingerprint leaves Papex, timestamped with OpenTimestamps and anchored in Bitcoin, with no wallet or cryptocurrency needed. You then keep it private, share it with people you invite, or publish it. Later edits become new versions with their own signature and date.

Each published version has a permanent address of the form papex.org/id/PX…/v/n and a citation to copy, as on Papex Labs' educational record PXH-0BADXVXY7QM3. If AI tools helped, the record can say which ones; model names are as you declare them (how to disclose AI use).

A Papex record shows who signed which exact version and that it existed by a date. It is not peer review or a preprint server's moderation, and it does not show that the work is correct, original or first. For how records, preprints, preregistration and timestamps compare, see how to protect a research idea before publishing.

Sources

• Attention Authors: updated endorsement policy (arXiv blog, 21 January 2026), checked 15 October 2026
• Attention Authors: updated endorsement policy for arXiv Mathematics (arXiv blog, 10 December 2025), checked 15 October 2026
• Endorsement (arXiv help), checked 15 October 2026
• How do I get endorsed? (arXiv User Support Knowledge Base), checked 15 October 2026
• How do I become an endorser? (arXiv User Support Knowledge Base), checked 15 October 2026
• Authority Records (arXiv help), checked 15 October 2026
• Example abstract page with the endorser link, arXiv:2501.00001 (arXiv), checked 15 October 2026
• Submission Guidelines (arXiv help), checked 15 October 2026
• Author Registration (arXiv help), checked 15 October 2026
• Identity and Affiliation (arXiv help), checked 15 October 2026
• arXiv moderation (arXiv help), checked 15 October 2026
• Policies for specific content types (arXiv help), checked 15 October 2026
• Attention Authors: Updated Practice for Review Articles and Position Papers in arXiv CS Category (arXiv blog, 31 October 2025), checked 15 October 2026
• Attention Authors: non-English Paper Submission Guidelines (arXiv blog, 13 January 2026), checked 15 October 2026
• Submissions in Languages Other Than English and Multiple-Language Submissions (arXiv help), checked 15 October 2026
• Fair Moderation, Equitable Access, and AI: arXiv's Updated Rate Limit Policy (arXiv blog, 1 October 2026), checked 15 October 2026
• Submission Status (arXiv help), checked 15 October 2026
• Appealing a moderation decision (arXiv help), checked 15 October 2026
• Affiliations (IOP Publishing, Publishing Support), checked 15 October 2026
• Taylor & Francis Editorial Policies (Taylor & Francis Author Services), checked 15 October 2026
• General Policies v1.0 (Zenodo), checked 15 October 2026
• About records (Zenodo help), checked 15 October 2026
• AI Usage Notice v1.0 (Zenodo, last revised 31 July 2026), checked 15 October 2026
• Welcome to OSF Preprints (OSF Support), checked 15 October 2026
• OSF Projects transition (OSF Support), checked 15 October 2026
• OSF Preprints (Center for Open Science), checked 15 October 2026
• What are SSRN's submission guidelines? (SSRN Support Center, Elsevier, updated 3 August 2026), checked 15 October 2026
• What is a DOI and how does it help my research? (SSRN Support Center, Elsevier), checked 15 October 2026
• SSRN home page (SSRN, Elsevier), checked 15 October 2026 (Internet Archive copy of 12 October 2026; the live page blocked automated access)
• Instructions for Authors (Preprints.org), checked 15 October 2026 (Internet Archive copy of 19 September 2026; the live page blocked automated access)
• About (Preprints.org), checked 15 October 2026 (Internet Archive copy of 19 September 2026; the live page blocked automated access)
• Editorial policies (Research Square), checked 15 October 2026
• Preprints (Research Square), checked 15 October 2026
• À propos de HAL (CCSD), checked 15 October 2026
• Documentation HAL (CCSD), checked 15 October 2026
• Auteurs et affiliations (HAL documentation, CCSD), checked 15 October 2026
• Vérification des dépôts avant leur mise en ligne (HAL documentation, CCSD), checked 15 October 2026
• About Qeios (Qeios), checked 15 October 2026
• Publishing policy (Qeios), checked 15 October 2026
• Frequently Asked Questions (bioRxiv), checked 15 October 2026
• About medRxiv (medRxiv), checked 15 October 2026 (Internet Archive copy of 1 October 2026; the live page blocked automated access)
• About ResearchGate (ResearchGate), checked 15 October 2026 (Internet Archive copy of 13 October 2026; the live page blocked automated access)