About one in three peer-reviewed papers published in 2024 came out under an open license, according to the most recent OurResearch counts on Unpaywall. The other two-thirds are still behind a paywall. If you left academia, work at a small biotech company, or no longer have alumni access privileges, you have probably hit the typical $40-$50 article paywall in the last month.

This guide walks through the legal options that exist in 2026: what each one is good for, what each one costs in time and money, and the rough threshold where each one stops scaling. Most independent researchers end up running three or four of these in parallel rather than picking one.

Who this is for

If you read scientific literature for a living without access to your university library, this guide is for you. That covers a wider population than it used to:

• Researchers who left academia and lost library access
• Academic researchers who have incubated a novel discovery, spun out of their university departments but have now exceeded their grace period
• Scientists at small biotechs, contract research organizations, and medical device startups
• Consultants, expert witnesses, and technical due-diligence analysts
• Medical writers, regulatory affairs contractors, and freelance pharmacovigilance specialists
• Nonprofit and policy researchers
• Researchers in low- and middle-income countries whose institutional subscriptions have lapsed or never existed

If you already have copyright permissible access to a strong university library or are employed at a large research institution, your problem is probably discovery, not access. This guide will be less useful for you.

The shape of the access problem in 2026

The 2026 landscape is genuinely different from the prior decade, driven by three shifts.

Open access grew. As of late 2024, OpenAlex put the share of open-licensed scholarly articles at about 50% of recent annual output, up from roughly 20% a decade earlier. The 2024 NIH Public Access Policy update, which took effect January 1, 2025, removed the 12-month embargo on NIH-funded research and pushed most biomedical work into PubMed Central immediately at publication. However, that NIH mandate (drafted by the OSTP and called the “Nelson Memo” named after the Deputy Director of the OSTP at that time) has since been put under review in 2026 and appears likely to be repealed, meaning paywall papers will once again be on the rise.

In parallel, Sci-Hub has stopped growing. Alexandra Elbakyan paused new uploads in late 2021 during ongoing litigation in India, and the corpus has been static since. In November 2024 the Delhi High Court ordered ISPs to block the main Sci-Hub domains, and the long-running ACS and Elsevier suit in the US has produced repeated takedown orders against mirrors. Sci-Hub raised real questions about pricing and access — questions that pushed publishers, funders, and platforms to move — but it is no longer a reliable place to find a 2024 paper.

Legal middle-ground tools matured. ReadCube, Sci-Net at libraries, DeepDyve, GetItNow, and the various Get Full Text Research browser extensions have made it possible to read most of the paywalled literature without an institutional contract.

With the NIH OA mandate likely to be repealed and with Sci-Hub now marginalized, researchers will need to consider new paths for accessing full-text papers. Fortunately, there are more legitimate routes than there used to be and the field is more cluttered. You have to know which route fits which situation.

The seven legal routes, compared

Route	Typical cost	Best for	Where it stops scaling
Open access (PMC, arXiv, bioRxiv, DOAJ, Unpaywall)	Free	First-pass searches, NIH-funded biomedical work, preprint-heavy fields	Misses most clinical, chemistry, materials, and humanities backfiles
Browser extensions (Unpaywall, Open Access Button, Get Full Text Research)	Free	Surfacing legal OA copies you’d otherwise miss	Only finds what’s actually open somewhere
Author email request	Free	Single high-value papers, older work, niche subfields	Slow and unreliable; useless under deadline
Alumni, public, hospital, or national library access	Free to ~$100/year	Occasional access, especially for life sciences	Coverage varies wildly; rarely covers your full reading list
Interlibrary loan (BLDSS, NLM DOCLINE)	$5-$30 per article	Specific papers you can wait 1-5 days for	Built for eventual delivery, not active research
Pay-per-view from publishers	$30-$50 per article	Single must-read papers you’ve already triaged	Costs run into hundreds of dollars per month fast
Subscription literature platforms (DeepDyve and similar)	$40-$60/month	Full-text, online reading across 25,000+ paywalled journals and ebooks	Coverage gaps in specific publishers; not a substitute for institutional access to everything

Each route has a frequency band where it works and a frequency band where it falls apart. That matters more than the headline prices.

Start with open access, but know what it doesn’t cover

Open access is the right first move every time, and it is incomplete.

The open layer in 2026 is much deeper than it was. PubMed Central holds over 10 million full-text biomedical articles. arXiv passed 2.5 million preprints in late 2024. bioRxiv and medRxiv together host more than 350,000. DOAJ indexes around 20,000 fully open-access journals. Unpaywall’s database covers roughly 50 million OA articles and the browser extension finds them automatically when you land on a paywalled page.

The gaps are predictable. Most Elsevier, Wiley, Springer Nature, and ACS subscription journals only have green OA versions for a fraction of recent articles, and those are often the accepted manuscript rather than the version of record. Chemistry, materials science, engineering, and most of the humanities lag biomedicine. Older literature — anything pre-2000 in many fields — is rarely open. Reviews and book chapters are usually paywalled even when the underlying research is open.

If your work is NIH-funded biomedical research from 2025 onward, open access will cover most of what you need. If you read clinical trial methodology, regulatory toxicology, or chemistry, plan for the open layer to cover maybe 30-40% of your reading list.

Author requests still work when you have time

Emailing an author for a copy is older than the modern paywall and still works. About half of polite requests get answered within a week in most fields. Senior authors with established labs reply more reliably than postdocs who have since moved twice.

It works best for one specific paper that matters to you, where the lead author is still active and easy to find. It fails when you need eight papers by Friday, when the author is no longer affiliated anywhere ResearchGate can route to, or when you are screening a hundred abstracts and won’t know which full texts you need until you read them.

Treat author emails as a fallback for individual high-value gaps, not as a workflow.

Library routes are real but slow

If you have alumni library access from a former university, hospital affiliation as a clinician, or borrowing rights at a national library, use them. They are legitimate and often free.

Three specific options most people overlook:

The New York Public Library’s SimplyE and research-card programs give NYC-area residents (and remote users with a library card) access to JSTOR and a meaningful slice of biomedical and humanities databases. The British Library’s On Demand service ships PDFs to non-affiliated users for £8-£15 per article with a copyright fee. The US National Library of Medicine’s DOCLINE network handles interlibrary loan for medical libraries, and many hospital librarians will help affiliated clinicians even when they’re not full staff.

Interlibrary loan usually runs one to five business days. That’s fine when you’re writing a literature review with a four-week deadline. It’s brutal when you’re trying to settle a methods question before a meeting at 3 p.m.

Pay-per-view: rational for one paper, punishing as a habit

Buying a paper directly from Elsevier, Wiley, Springer Nature, or Taylor & Francis is the fastest legal route when you already know you need that exact article. Prices in 2026 typically run $30-$50 per paper. Wiley and Elsevier hover around $39.95-$49.95 for most journals. Some Springer Nature titles charge $59.

At modest volume the math stops working. Five papers a week at $40 each is $10,400 a year. Most independent researchers who track their pay-per-view spend for a quarter end up surprised by it, and they often realize a chunk of those purchases were exploratory reads where the abstract turned out to be misleading.

There is also a behavioral cost. When every full text costs $40, people stop checking the methods section. They stop reading the supplementary material. They work from abstracts longer than they should, and they miss the negative findings, the subgroup data, and the limitations that live in the parts of the paper they didn’t open.

Sometimes the abstract is enough. For most research questions it isn’t.

What about Sci-Hub?

Sci-Hub raised legitimate questions about scholarly publishing pricing and access — questions that genuinely changed how publishers, funders, and platforms behave. As a practical option for current research, it no longer works.

Alexandra Elbakyan paused new uploads in December 2021 during the Indian copyright suit and the corpus has not been updated since. Anything published after 2021 is unlikely to be there. The main domains face ongoing takedown orders, the Delhi High Court blocked Indian ISP access in November 2024, and the US lawsuit from ACS continues to produce mirror seizures.

The legal middle ground covers most use cases now: open access for what’s open, library routes for what your affiliation covers, and a subscription platform when paywalled reading is part of your week. In 2017 that combination didn’t hold together. In 2026 it does.

When a subscription literature platform makes sense

Once paywalled reading becomes a weekly habit rather than a monthly inconvenience, the per-article model stops working. Subscription literature platforms exist for that frequency band.

DeepDyve serves independent researchers and small teams who don’t have an R1 library and don’t have a six-figure budget for institutional contracts. The Personal plan is $59/month or $499/year, and it includes full-text streaming access to journals from Springer Nature, Wiley, Taylor & Francis, SAGE, Oxford University Press, and roughly 100 other publishers. The collection covers around 35 million full-text articles in the streaming layer, plus open access discovery across about 150 million records.

Specifics worth knowing before you decide it fits:

Elsevier is not in the DeepDyve streaming collection. If your research needs are heavily concentrated in The Lancet or Cell Press, for example, that’s a real gap and you should check your reading list against the journal list before subscribing. ACS journals are also outside the streaming collection. However, DeepDyve does include every journal from Springer Nature, Wiley, Sage, JAMA, NEJM, and many other titles from nearly all of the top scientific publishers.

The streaming model means you read the article in the browser; you don’t download a PDF for offline storage. That works for most people and frustrates a minority who want to print and mark up PDFs to save as a hard copy.

Enterprise plans add shared archives, duplicate-purchase prevention, and copyright-cleared business use, which matters if you’re at a biotech or high tech company using the literature for commercial work rather than personal study.

DeepDyve isn’t the only option in this band. ReadCube Papers and Reprints Desk all play in adjacent territory with although their offerings are limited to transactional pricing where you pay per article, whereas DeepDyve offers the all-you-can-read end option, as well as pay per use.

A realistic stack

Most independent researchers end up combining tools rather than picking one. A common strategy might be:

1. Open access tools (Unpaywall extension in the browser, PubMed Central searches, arXiv alerts) handle the first 20-30% of what they need to read.
2. A subscription platform of aggregated titles such as DeepDyve handles full-text reading across paywalls journals they touch most.
3. Pay-per-article for any papers missing from the subscription, or pre-pay for article tokens from the one punisher or journal that is outside the subscription
4. Lastly, contact authors directly or use interlibrary loans to avoid or minimize article purchases, however be aware that there may be lengthy time delays while waiting for your request.

Each route gets used at the frequency band where it actually works. That’s what makes the stack hold together.

How to decide what fits your situation

If you read one or two papers a month from outside what’s openly available, pay-per-view or author emails are fine. The math doesn’t justify a subscription.

If you read five or more paywalled papers a month and your work is in fields with substantial open access (most NIH-funded biomedicine, machine learning, physics, mathematics), see how far Unpaywall, PMC, and arXiv get you before paying for anything.

If you read paywalled papers most weeks, work in a field where open access lags (chemistry, materials, engineering, clinical research outside NIH funding, regulatory toxicology), or are part of a team where multiple people are buying the same articles, a subscription platform is usually the cheapest answer per article read.

If you only need access to one specific publisher’s journals — for example, you read almost entirely Elsevier titles — a single-publisher pay-per-view account or a ReadCube tokens setup may beat a multi-publisher subscription. Check the math against your last three months of reading.

FAQ

How can I legally access research papers without a university login? Start with open access sources — PubMed Central, arXiv, bioRxiv, DOAJ, and the Unpaywall browser extension. For paywalled papers you need occasionally, use author email requests, alumni or public library access where you have it, or interlibrary loan through services like the British Library On Demand or DOCLINE. For recurring full-text reading, a subscription platform like DeepDyve or a transactional account with the publishers you read most.

What percentage of papers are open access in 2026? OpenAlex and OurResearch put the share of recent peer-reviewed articles with some form of open license at roughly 50% as of 2024, with coverage skewed toward biomedicine (higher) and chemistry, engineering, and humanities (lower). The 2024 NIH Public Access Policy update made most NIH-funded biomedical work immediately open in PubMed Central from January 2025.

Does pay-per-view ever make sense? Yes — for one paper you already know you need. Publisher prices typically run $30-$50 per article. Above about three to five paywalled papers a month, the math usually favors a subscription platform.

Is Sci-Hub still working? The Sci-Hub corpus has not been updated with new uploads since late 2021. Several jurisdictions, including India in 2024, have ordered ISPs to block its main domains, and the US copyright suit from ACS continues to produce takedown orders against mirrors. For current research, the legal alternatives have caught up enough to be the practical answer.

When does DeepDyve make sense for an independent researcher? When you read paywalled papers most weeks, when your reading isn’t heavily concentrated in Elsevier or ACS titles (which aren’t in the streaming collection), and when you’d rather pay $59/month than build the same access workaround for every paper. The Personal plan covers independent researchers, students, and hobbyists.

What’s the best option for a small biotech or CRO team? Once multiple people on a team need recurring full-text access, scattered individual pay-per-view purchases get expensive and duplicative quickly. A Team plan on a subscription platform, a managed document-delivery account with Reprints Desk or Article Galaxy, or a small institutional contract through a publisher are the main legitimate options. For commercial use specifically, you also need copyright-cleared access — some routes that work for individuals do not cover commercial use.

Does DeepDyve include Elsevier or ACS journals? No. Elsevier and ACS journals are not in DeepDyve’s streaming collection. If your reading list is heavily concentrated in those publishers — for example, you mostly read The Lancet, Cell, or JACS — you should check the journal coverage against your reading list before subscribing.

Give it a Try

If you read paywalled papers most weeks and you’re tired of solving the same access problem one paper at a time, start a 30-day free trial of DeepDyve and check the journal list against your last month of reading.

DeepDyve pricing
DeepDyve features
Browse journals on DeepDyve