How to Download Research Papers for Free: 7 Legal Sources

April 23, 2026

Reviewed

Paywalls are the single biggest friction point in research. If you are wondering how to download research papers for free without breaking the law or losing half a day to dead ends, this guide covers seven legal routes that work in 2026 — from PubMed Central and preprint servers to Unpaywall, CORE, and direct author sharing. It is aimed at graduate students, independent researchers, clinicians, and anyone outside a well-funded institution who still needs access to full-text evidence. Every source below is legal; we also briefly address why Sci-Hub is not a safe option. Work through the list in order — most readers find their paper within the first three.

1. PubMed Central (PMC)

PubMed Central is the NIH's free full-text archive of biomedical literature. If a paper was funded by the NIH, Wellcome Trust, or any funder with an open-access mandate, a full-text copy is almost certainly deposited in PMC within 12 months of publication. Start every biomedical search here.

Workflow: go to pubmed.ncbi.nlm.nih.gov, run your query, and tick the Free full text filter in the left sidebar. Every hit with a PMC icon can be read or downloaded as PDF or XML at no cost. PMC also hosts thousands of full books and monographs via NCBI Bookshelf. For clinicians, the NIH's LitCovid subset and author-manuscript deposits cover a huge swathe of post-2008 biomedical output.

2. Preprint servers (bioRxiv, medRxiv, arXiv, ChemRxiv)

A preprint is the author's manuscript before peer review. For roughly 60% of biomedical papers published since 2020, an open preprint exists on bioRxiv or medRxiv that is textually near-identical to the final version. Preprints are free, fully legal, and posted by the authors themselves.

  • bioRxiv — life-science preprints (Cold Spring Harbor Laboratory)
  • medRxiv — health-science and clinical research
  • arXiv — physics, maths, computer science, quantitative biology
  • ChemRxiv — chemistry

Each server links the preprint to its peer-reviewed version once published. For tracking this correspondence automatically, see Unpaywall below.

3. Unpaywall (browser extension)

Unpaywall is a free browser extension that scans over 50,000 journals and repositories for a legally open version of whatever paper you are looking at. Install it, land on a paywalled article page, and a green padlock icon appears in the corner if a free version exists elsewhere (institutional repository, preprint, author page). One click opens the PDF.

Unpaywall only surfaces copies that authors or publishers have legitimately made open — it does not scrape pirated sources. It covers roughly 50% of all recent biomedical articles. Pair it with a second click-through to CORE if Unpaywall does not find a match.

4. CORE and OpenAIRE

CORE is the world's largest aggregator of open-access research, indexing over 300 million records from more than 11,000 repositories. Where Unpaywall returns a single best match, CORE lets you browse every known open copy (institutional repositories, thesis archives, preprint servers) and download the one that suits. OpenAIRE does the same for European funded research.

Use CORE when Unpaywall comes up empty or when you want to audit multiple versions (e.g. compare the accepted manuscript against the published version of record).

5. Email the corresponding author

Every published paper lists a corresponding author and email. Authors are generally delighted to receive reprint requests and usually reply within a day or two with a PDF attached. This is legal, encouraged, and how researchers shared work for decades before the web.

Template: one short paragraph — who you are, which paper you are requesting (title + DOI), one sentence on why it is relevant to your work. No need to justify at length. Many authors also post reprints automatically on ResearchGate or Academia.edu; a single search there often saves the email.

6. Institutional and national open-access platforms

Many universities host their entire institutional output in a free repository: Harvard DASH, MIT DSpace, Oxford ORA, UCL Discovery, and hundreds of others. National platforms (HAL in France, PUMA in Italy, DiVA in Scandinavia) aggregate thousands of universities at once. These hold green open-access versions — author manuscripts that are legally free even when the publisher version is paywalled.

Your own institution, if you have one, should have negotiated transformative agreements with major publishers (Elsevier, Wiley, Springer Nature) that give staff and students full-text access. Check your library's A-Z journal list before paying.

7. Interlibrary loan (ILL) and national libraries

Public libraries in most countries run interlibrary loan services that will source any paper in days, usually free of charge or for a nominal fee. The British Library, US National Library of Medicine, and Germany's TIB act as national backstops — between them they hold almost everything. For rare, old, or non-English papers that none of the above routes surface, ILL is the definitive answer.

What about Sci-Hub?

Sci-Hub hosts tens of millions of paywalled papers without publisher permission. It is popular, but using it carries real risks: accessing papers through Sci-Hub violates publisher copyright in most jurisdictions, some institutions block Sci-Hub traffic and log attempts, and the ethical argument against it is not settled. The seven legal routes above cover 80–95% of biomedical literature. Exhaust them before considering anything else.

Common mistakes

  • Googling the title once and giving up. Try Google Scholar, then PubMed, then Unpaywall, then CORE — in that order. Most "paywalled" papers have a free copy within two clicks if you know where to look.
  • Ignoring preprints because they are "not peer-reviewed". For methods and results, preprints and final versions are usually textually near-identical. Cite the final version once published; read the preprint in the meantime.
  • Downloading pirated copies through unfamiliar domains. Beyond the legal issue, mirror sites can serve malware-laden PDFs. Stick to the sources above.
  • Forgetting your own library. If you are enrolled anywhere, check the library proxy first — you may already have access.
  • Paying $35 per paper. Never do this until every route above has failed. For difficult cases, one £3 interlibrary loan through your public library is the cheapest last resort.

Tools and resources

  • BioSkepsis — a biomedical AI research assistant with a biology-native knowledge graph over 40M+ curated papers; prioritises open-access sources during search and shows full-text evidence inline.
  • PubMed / PMC — always the first stop for life-science literature.
  • Unpaywall — browser extension; the single biggest time-saver for legal OA access.
  • Google Scholar — broad coverage, shows "[PDF]" links to OA copies in the right-hand column.
  • CORE — when you need to compare multiple open versions.
  • Semantic Scholar — 200M+ papers with an open API, useful for programmatic retrieval.

How BioSkepsis helps

When BioSkepsis surfaces a paper in its search results, it also flags whether a legally open full-text version exists and routes directly to PMC or the preprint server where possible. Because the platform reasons over full text (methods, controls, supplementary) rather than abstracts alone, researchers spend less time hunting for PDFs and more time evaluating evidence. The free tier covers 100 papers per session with Zotero sync built in.

Frequently asked questions

Is it legal to download research papers for free?

Yes, when you use open-access sources. PubMed Central, preprint servers, Unpaywall, CORE, author-shared copies on personal pages, and interlibrary loan are all fully legal. What is not legal in most jurisdictions is accessing paywalled publisher content through unauthorised mirror sites.

What percentage of research papers are available for free?

Roughly 50–60% of biomedical papers published since 2018 have a legally open full-text version somewhere — as a preprint, a green OA deposit, or a gold OA publication. For NIH-funded research the figure is closer to 95% within 12 months of publication.

Can I download papers from Google Scholar?

Indirectly, yes. Google Scholar indexes both paywalled and open versions. When an open PDF exists, Google Scholar shows it as a "[PDF]" link in the right-hand column of the search result. Click that link to download directly from the hosting repository.

How do I find free research papers without an institutional login?

Start with PubMed Central's Free full text filter, install the Unpaywall browser extension, and keep Google Scholar and CORE as fallbacks. Between those four, most recent biomedical literature is accessible.

Should I use ResearchGate to request papers?

ResearchGate's "Request full-text" button routes directly to the author. It works fine, but has no advantage over a plain email and some papers there have been posted in violation of publisher agreements. A direct author email is cleaner.

Try BioSkepsis free — no credit card

Biology-native knowledge graph across 40M+ biomedical papers. Free tier with 100 papers per session, Zotero sync, full-text reasoning — with open-access sources flagged inline.

Start free