Zotero — Free Reference Manager for Researchers

What Zotero is

Zotero is a free, open-source reference manager that captures papers from your browser, stores their metadata and PDFs, and inserts formatted citations into whatever you write in. It is built by the Corporation for Digital Scholarship, a non-profit that grew out of the Roy Rosenzweig Center at George Mason University.

The architectural fact that explains everything else: Zotero is a local application with an optional sync layer, not a cloud service with a desktop skin. Your library is a SQLite database and a folder of files on your own disk. Sync is a convenience you can switch off, replace with your own WebDAV server, or lose entirely without losing your references. Nothing else in this category can say that, and it is why Zotero is the default recommendation on this site.

Why researchers use it

• One-click capture — the browser connector reads the page, pulls the metadata, downloads the PDF, and files it. Journal pages, arXiv, PubMed, and library catalogues all work.
• Citation Style Language — over 10,000 journal styles in the public repository, so “reformat for the other journal” is a dropdown rather than an evening.
• Better BibTeX — the plugin that gives you stable, human-readable citation keys (zeghlache2024arxiv) and auto-exports a .bib file your LaTeX project can point at forever.
• Writes where you write — Word, LibreOffice, and Google Docs plugins officially; Markdown, LaTeX, Obsidian, and Quarto through the .bib export.
• Group libraries — shared collections for a lab or a systematic review, free and unlimited for public groups.

Where it fits in a research workflow

Zotero sits at the front of the writing pipeline and feeds almost everything downstream. Capture happens in the browser; annotation happens in Zotero’s built-in PDF reader; the output is either a formatted citation (Word, Google Docs) or a .bib file.

That .bib file is the interesting boundary. Point Overleaf at it and your LaTeX bibliography maintains itself. Point Pandoc at it and you can write in Markdown with [@zeghlache2024arxiv] and export to Word, PDF, or HTML with a correct bibliography in any style — the workflow in Write Research Papers in Markdown + Pandoc. Point Obsidian at it and your literature notes carry live metadata. One capture, every destination.

Getting started

Ten minutes to working, an afternoon to a setup you will not rebuild.

1. Install the desktop app and the browser connector — the connector is where the time saving actually lives. Capture ten papers from your current project before configuring anything.
2. Install Better BibTeX from its releases page. Set a citation-key format and leave it alone; changing key formats after you have cited 200 papers is the one genuinely painful mistake in Zotero.
3. Turn on Settings → Export → Better BibTeX → Automatic export so a .bib file stays in sync with a collection on disk.
4. The step everyone skips: set your file-storage folder deliberately, then check it is inside whatever your institution backs up. Zotero 7 renames and files attachments natively — you no longer need ZotFile for this.

Zotero vs the alternatives

Pick Zotero unless you have a specific reason not to. The reasons that hold up are an institutional mandate, a LaTeX-purist preference for JabRef, or a Google Docs–centric writing group.

Cost, licensing, and your data

The software is free and AGPL-licensed with no feature gating — paid plans buy storage, nothing else. The free sync tier is 300 MB, which is roughly a few hundred annotated PDFs and which a serious PhD library will exhaust. Two escapes: pay for a storage tier, or point Zotero at your own WebDAV endpoint (many universities provide one) and sync unlimited attachments outside Zotero’s servers entirely.

Your library lives on your machine by default. For work under an embargo or a data agreement, that local-first default plus WebDAV means you can run Zotero without any third-party service ever holding your files.

The honest review

Strengths. Zotero is the only reference manager where you are not a tenant. The data format is open, the export is complete, the plugin API is real, and a non-profit with a two-decade track record maintains it. Over a career, that ownership compounds into something no interface polish can match.

Limitations. The interface is functional rather than pleasant, and it shows its age against Paperpile. Metadata capture from publisher pages is good but not perfect — expect to fix author fields and journal abbreviations by hand, especially for older or non-English sources. Better BibTeX is close to mandatory for LaTeX users yet is a community plugin, not core, so a major Zotero release can briefly break it. Group-library sync is slower than the commercial competition.

Verdict. Adopt it as your default and stop shopping. Skip it only if your institution mandates something else or your entire writing life happens inside Google Docs. The condition that flips the answer is collaboration: if your co-authors will not install it, the friction lands on you.

Common questions

Is Zotero really free?

Yes. The application is free and open source with every feature unlocked, forever. The only paid element is cloud storage above the 300 MB free tier, and you can avoid even that by syncing attachments through your own WebDAV server while metadata sync stays free.

Does Zotero work with LaTeX and Overleaf?

Yes, through BibTeX export. Install Better BibTeX, set a collection to auto-export as a .bib file, and point your LaTeX preamble at it. For Overleaf specifically, you can also link a Zotero library directly from the project’s bibliography menu.

Can I move my library out of Zotero later?

Yes, and this is the point. Everything exports to standard formats — BibTeX, RIS, CSL JSON — and your PDFs are ordinary files in a folder you chose. Migration to another manager is a straightforward import, not a rescue operation.