Zotero vs Mendeley in 2026: Which Reference Manager Should Researchers Use?

Most researchers choose a reference manager the way they choose a dentist: someone in the lab recommended it, it worked well enough, and nobody revisited the decision. That can last for years. Then a manuscript is rejected, transferred to a second journal with a different reference style, and the bibliography has to be rebuilt by hand at eleven at night. The choice between Zotero and Mendeley is one of the few tooling decisions in academic writing that quietly shapes every submission you make afterward.

The two tools are the most common free options for researchers who want something better than a spreadsheet of DOIs. Both can capture references from websites and databases, store PDFs, sync across devices, and insert citations into Microsoft Word. Both are used heavily in clinical research. The differences are real, though, and they lie in ownership, storage, the way each handles citation styles, and how comfortable each is for collaborative groups. This guide compares them with the medical manuscript in mind.

Why the Choice Matters More in Medical Publishing

A clinical paper rarely stays with one journal. A trial report might be declined by a general medical journal, sent to a specialty title, and then moved to a sister journal within the same publisher. Each step can change the reference style, the reference limit, and the way in-text citations are numbered. Vancouver style, the numbered format used by most biomedical journals and based on the reference examples in the ICMJE Recommendations, has many small variants. Some journals want abbreviated journal names from the NLM catalog. Others want full names. Some cap the number of authors listed before et al., and some want DOIs appended.

Doing those conversions by hand is slow, and it is where errors creep in. A reference that reads correctly in one journal can end up with a mangled author list or a missing volume number after a manual reformat. In the worst case, a citation that points to the wrong paper survives into the accepted version. Reference managers do not eliminate that risk, but they make the conversion mechanical, which means a single style change can be checked once rather than dozens of times.

What Each Tool Is, and Who Runs It

Zotero is open source and developed by the Corporation for Digital Scholarship, a nonprofit. Its code is publicly available, its desktop application is free, and it does not depend on a single company's business model to keep running. Its browser connector captures references and PDFs from most library catalogs and publisher sites with one click. Recent releases have focused on reading and library management, and the 2026 release announcements for Zotero 9 and Zotero 10 describe features such as a built-in reader with read-aloud, annotation integrations and improved search. Check the release notes for the version you install, because these details change faster than most guides are updated.

Mendeley has belonged to Elsevier since 2013. That ownership matters for two reasons. First, it means the tool is tightly connected to one of the largest journal publishers in the world, which is an advantage for some workflows and a concern for researchers who care about data governance. Second, Mendeley's roadmap is set by a commercial organization whose priorities may not match those of individual researchers. Neither point is a verdict on the product. Both are facts you should weigh when you commit several years of references and PDFs to one system.

Several 2026 comparison guides, many of them written by vendors of competing tools, describe Mendeley as having lost ground to Zotero among researchers who prioritize openness. Those claims are worth treating with some caution. They come from commercial sources with an interest in the outcome, and I could not locate an official Elsevier statement about Mendeley's direction in the sources I reviewed. What you can verify directly is the current feature list, the pricing page, and the data export options on each tool's own website.

Cost and Storage: Where the Numbers Stand

Both tools have a free tier, and the distinction that matters most is what the free tier limits. Zotero syncs your library metadata for free but limits file storage for PDFs and attachments, with a default allowance of 300 MB at the time of writing. Mendeley's free plan has generally offered more cloud storage, with most sources I checked putting it at 2 GB. Anyone with a large PDF collection will hit the Zotero limit quickly, and a clinical researcher who reads trial reports, guidelines and supplementary files can accumulate several gigabytes within a year.

Paid storage is sold by both, and this is where comparison articles disagree most. Prices differ by region, change over time, and are sometimes bundled differently. The sources I found quoted incompatible figures for Mendeley's paid plans, so I will not repeat any of them here. Before you choose, open each tool's current pricing page, check the storage you would need for your existing PDFs, and ask whether your institution already provides a license or a group library. Many universities do, and a departmental plan can change the answer entirely.

Citation Styles and Switching Journals

Both tools are built on the Citation Style Language, an open standard that describes how references should be formatted. Zotero advertises more than 9,000 styles, and Mendeley offers a large catalog as well. Vancouver is available in both, along with the common alternatives that biomedical journals use. For most medical authors, the practical question is not whether a style exists but how easily you can change to it and whether the output matches the journal's instructions line by line.

Here the choice of tool matters less than the discipline around it. A style file that is almost right still needs checking. Look at how the journal title is abbreviated, how many authors appear before et al., whether issue numbers are shown, and how the DOI is formatted. Make the check once for each target journal and save the result as a named style. The useful habit is not to trust the first render. It is to compare two or three references against the journal's own examples in its author instructions, because those examples are the ones the production team will follow.

Transfers are where this pays off. If a manuscript moves from one journal to another after review, the reference list can be regenerated in the new style without touching the text. Authors who keep their citations as plain typed numbers have no such option. The same logic applies to the journal-to-journal cascade that has become common in clinical publishing, a process covered in our guide to manuscript transfers. A library that is already organized by reference style makes that move a short task rather than a bibliography rebuild.

Word, LibreOffice, Google Docs and LaTeX

Most clinical co-authors still work in Microsoft Word, and both tools offer a plugin for it. Zotero's Word plugin inserts citations, builds the bibliography, and can switch styles across the whole document. Mendeley's Cite plugin does the same in its own way. Both plugins have occasionally broken after Word updates, so it is worth testing the plugin on a short document before a deadline. If a co-author uses LibreOffice or Google Docs, check the current plugin support for each tool before you commit a team to one setup, because plugin coverage changes more often than the core features do.

LaTeX users have a different calculation. Both tools can export BibTeX, and for many researchers a reference manager is only a bibliography source feeding a .bib file. Zotero's large community of plugins includes tools that improve BibTeX key generation and keep the file synchronized, and the ecosystem around Zotero is generally deeper for LaTeX workflows. If your group writes in Overleaf, export a fresh .bib file from whichever tool you choose and confirm that the keys match your citation commands. A mismatched key produces a question mark in the PDF, and an editor is unlikely to be impressed by one.

Annotation, Groups and Shared Libraries

Both tools let you highlight PDFs and attach notes, and for a single researcher the difference is mostly a matter of interface preference. Zotero's reader is part of the desktop application and has been expanding in recent releases. Mendeley's annotation tools are mature and familiar to many users who came to it as students, and some researchers find them more pleasant for light reading.

Collaboration is where the decision can tip. Mendeley's group features have long been a selling point for research teams, and its shared group libraries are simple to set up. Zotero supports group libraries as well, with the option to make them public or private and to assign members different permissions. For a clinical trial group with several sites, the practical question is who needs to add references, who needs to edit them, and whether the group library will still be accessible when a postdoctoral researcher leaves. Write that policy down before the group library grows.

Moving a Library from Mendeley to Zotero

Zotero includes an importer for Mendeley libraries, which brings across references and folder structure, and in many versions some annotations as well. Migration is usually straightforward for a library of a few hundred references. For a larger collection, the time-consuming part is checking what survived. Some metadata fields map imperfectly, and attachments may need to be reconnected. Do not delete the Mendeley library until you have checked a sample of references, including items with unusual author names, non-English titles and missing DOIs.

Going the other direction is less common, but it is possible through standard exports. Whichever way you move, export a backup in a neutral format such as BibTeX or RIS before you start, and keep it. If the migration fails halfway through a revision, the backup is the difference between an annoying afternoon and a lost bibliography. Most researchers never think about this until it is too late, which is why it belongs in the checklist rather than in an emergency plan.

What a Reference Manager Cannot Do

Neither tool checks whether a reference is real, accurate or still valid. A reference manager will happily format a citation that was copied from a generative AI tool, including one for a paper that does not exist. One recent analysis of PubMed-indexed papers found that roughly one in 277 papers published in the first seven weeks of 2026 cited a paper that does not exist, and editors have good reason to ask authors to check every citation. Reference managers can help with the verification step, since their metadata comes from bibliographic databases, but they cannot replace a human checking the title, authors and DOI against the source.

Retraction status is a related gap. Zotero has in recent versions flagged items that appear in retraction databases, and it is worth confirming how your version handles this. Treat any flag as a prompt to check, not as a complete screen. Retractions are sometimes added after a library entry is saved, and a manager's metadata may lag behind the publisher's notice. Our guides to checking retracted citations and to AI-fabricated references cover the manual checks that should happen before submission, whichever tool you use.

How to Choose: A Practical Decision Path

If you work alone, write in Word, mostly submit to medical journals, and care about long-term openness, Zotero is the more defensible default. It is free, open source, widely supported, and easy to move your data out of. If your group already uses Mendeley groups heavily and your institution pays for an Elsevier license, staying put is reasonable, especially if the group's shared library works well. If you write primarily in LaTeX, Zotero's ecosystem is usually the better fit, and the BibTeX workflow is easier to automate.

The decision is less important than the discipline around it. Pick one tool for your whole group, agree on how references are named and tagged, and keep the library organized by project rather than by vague categories like reading or misc. A tidy library survives a change of tool. A messy one tends to stay messy in whichever application you move to.

What to Do This Week

If you are happy with your current tool, spend an hour on the setup rather than the software. Create one named Vancouver style for your most likely target journal, test it on five references, and export a backup. If you are not happy, or you are starting a new project, run a small trial: import twenty references from a finished paper, format them in two journal styles, insert them into a Word document, and open the file on a second computer. If all of that works, you have made a better decision than most authors who simply installed whatever a colleague suggested.

The practical rule is simple. Your reference manager should make the next journal switch boring. If it does not, the tool is not the problem, and it is worth finding out why before the next manuscript is rejected.

Further Reading