How to Find a Free, Legal PDF of a Research Paper

You find the exact paper you need, click through, and hit a $39.95 paywall for a PDF that is eight pages long.

Before you give up or pay: for a large share of published research there is a legal, free copy sitting somewhere else. Not a pirated one — a version the author or publisher is entitled to give away. The trick is knowing where those live.

Start with the DOI

Every published paper has a DOI — the identifier that looks like 10.1038/s41586-021-03819-2, usually printed near the title or in the citation.

The fastest route is to check the DOI against an index of known open-access copies. That is what Unpaywall does: a database of legal free versions, harvested from publishers and repositories, covering tens of millions of articles.

Our Paper Finder queries it directly. Paste a DOI — or just the paper's title, which it resolves to a DOI through Crossref first — and it returns the legal free copy if one exists. It runs entirely in your browser, needs no account, and if there is no legal copy it tells you so rather than sending you somewhere dubious.

That single check resolves most cases. When it does not, work down this list.

Where legal free copies actually live

Preprint servers. Many authors post the paper before peer review. arXiv covers physics, maths, computer science and increasingly other fields; bioRxiv and medRxiv cover life sciences and medicine; SSRN covers social sciences and law; RePEc covers economics. These are fully legal and usually free to read forever.

PubMed Central. Research funded by the NIH, Wellcome, and many other public funders must be deposited in a public archive. If the work is biomedical and publicly funded, PMC or Europe PMC very often has it.

Institutional repositories. Most universities run an archive of their own researchers' output, and most publishers permit authors to deposit the accepted manuscript there — this is what "green open access" means. Searching the author's surname plus their university plus "repository" finds these surprisingly often.

The author's own website. Academics routinely post PDFs of their papers on a personal or lab page. A search for the exact title in quotation marks, plus filetype:pdf, turns these up.

Google Scholar. Scholar shows a [PDF] link to the right of a result when it knows of a free copy. It is not exhaustive, but it costs one search.

Just ask the author

This is the most underrated route and it works far more often than people expect.

Authors are usually delighted that someone wants to read their work, and sharing a copy for individual scholarly use is normally within their rights. Email addresses are in the paper itself, or on the department page.

A short message is enough:

Dear Dr Okonkwo,

I am a master's student at [university] working on [topic]. I would very much like to read your paper "[exact title]" but do not have institutional access to [journal]. Would you be willing to share a copy?

Thank you for your time.

Expect a reply within a few days more often than not.

Your library, even if you are not a student

If you are affiliated with a university, log in through the library's proxy before you conclude a paper is locked — access often looks paywalled until you authenticate.

If you are not affiliated, interlibrary loan still exists and most public library systems offer it. You request the article, a library that holds it sends a copy, and it is free or close to it. It takes days rather than seconds, but it is a genuine route that most people have forgotten about.

A note on ResearchGate and Academia.edu

You will often find the PDF you want on one of these. Be aware that uploads there are frequently made without the rights to do so, and publishers periodically issue takedowns — copies vanish, and links rot.

If the same paper is in an institutional repository or on a preprint server, prefer that copy. It is unambiguously legal and it will still be there in five years, which matters if you are citing it.

About the shadow libraries

Sites that host pirated copies of paywalled papers are well known, and we do not link to them. They distribute copyrighted work without permission, which is unlawful in most countries regardless of how reasonable the frustration behind them is.

The practical point is that you usually do not need them. Between preprints, funder mandates, and green open access, a large fraction of recent literature has a legal free version — and the legal copy is the one you can cite without wondering whether the link will survive.

Which version are you reading?

Free copies come in three flavours, and it matters for citation:

• Preprint — before peer review. The findings may have changed substantially in the published version. Cite it as a preprint, not as the article.
• Accepted manuscript — peer-reviewed and accepted, but not yet typeset by the publisher. The content matches the published paper; the page numbers do not.
• Version of record — the final published article, with the journal's formatting and pagination.

If you are quoting with page numbers, you want the version of record, or you need to make clear which version you used. If you are reading to understand the work, the accepted manuscript is identical in substance.

Our Citation Generator will build the reference from the DOI, which resolves to the published version — but check the version you actually read before quoting a page number from it.

The short version

Try the DOI in an open-access finder first. If that fails, check preprint servers and PubMed Central, then the author's repository or personal page, then email them. Interlibrary loan is the backstop. Somewhere in that list is a copy you can read without paying $39.95 or breaking the law.