5.0 · 500+ projects delivered

PubMed to BibTeX Converter

Free. No signup. Runs in your browser.

Paste a PMID, get clean BibTeX. Page ranges are expanded (436-44 becomes 436--444), preprint IDs don’t leak into the pages field, full journal names come through instead of NLM abbreviations, and acronyms in titles stay capitalised – ready to drop into your .bib without breaking the compile.

How to convert PMIDs to a BibTeX file online

Visit the PubMed to BibTeX tool page You’re already here. Bookmark thelatexlab.com/pubmed-to-bibtex/ for next time.
Step 1
Paste your PMID One PMID like 26017442, or up to 50 – one per line – for batch conversion. Pure numeric IDs only; PMCIDs (with the PMC prefix) belong on a different path. If you only have a DOI, use /doi-to-bibtex/ instead. If your references live inside a Word manuscript, the references-conversion guide covers the path from EndNote / Mendeley in Word to a clean .bib.
Step 2
Pick the output dialect BibLaTeX preserves UTF-8 for biber. Legacy BibTeX rewrites accents as LaTeX macros (Kr{\"a}mer) for older bibtex setups. Default is BibLaTeX.
Step 3
Convert, then copy or download Each entry renders with its inferred type and a per-row entry-type override. Use the per-row Copy button, or Copy all / Download .bib at the bottom for the full batch.
Step 4
Lookups happen in your browser – PMIDs go straight from your browser to Europe PMC’s REST API, never to our server. We use Europe PMC instead of NCBI E-utilities because Europe PMC sends open CORS headers (same underlying MEDLINE data, browser-direct). Entries that fail to resolve show up inline with the reason so you can fix or drop them without losing the rest of the batch.

Where most PubMed converters break

Preprint IDs leak into the pages field For COVID-era papers that started life on medRxiv or bioRxiv, Europe PMC’s pageInfo field sometimes contains the preprint ID itself (e.g. 2020.03.27.20044925) instead of a real page range. Most converters pass that straight through as pages = {2020.03.27.20044925}, which compiles but produces a nonsense citation. The tool detects multi-dot values and skips the field rather than emit garbage.
filtered non-page values dropped
Compressed page ranges stay compressed NLM’s convention is to truncate the end page when digits repeat: 436-44 means pages 436 through 444, not 36 through 44. Most converters emit the literal pages = {436-44}, which is technically valid BibTeX but renders as the wrong range under any bibstyle that doesn’t expand it. The tool expands compressed ranges to 436--444 when the result reads forward.
436--444 expanded at conversion
Acronyms in titles get lowercased Biomedical titles are dense with acronyms – DNA, RNA, mRNA, COVID-19, SARS-CoV-2, HIV, IL-6. Under plain, abbrv, and most bibstyles that case-fold titles, an unprotected title turns into “Dna methylation in covid-19 patients”. The portable fix is wrapping all-caps tokens in braces: {DNA}, {COVID-19}. If you’ve ever gotten a “Missing } inserted” after pasting a title with unbalanced braces, that’s the symptom the protection step prevents on the way in.
{DNA} applied at conversion
Entry types don’t distinguish reviews and letters Europe PMC tags every MEDLINE record as a journal article, including review articles, letters to the editor, editorials, and case reports. Bibstyles render these differently when typed correctly. The tool uses @article by default and exposes a per-row entry-type override so you can switch a single entry to @incollection, @misc, or whatever fits before copying.
Override per-result dropdown
The journal field comes from Europe PMC’s full journal title rather than the NLM-abbreviated form (New England Journal of Medicine, not N Engl J Med). When the PubMed record carries a DOI, it’s included alongside the PMID so the entry resolves under either path. Citekeys are lastnameYearWord with collision handling.

Frequently asked questions

If the paper has a DOI (most journal articles published after 2000 do), cite by DOI – it survives if PubMed reorganises its indices and resolves outside biomed contexts. Cite by PMID when the paper predates DOIs or when your style guide explicitly requires it (some clinical journals do). PMCIDs identify the full-text deposit, not the article itself, so they belong in pmcid or a note field rather than as the primary identifier. The tool returns both PMID and DOI in the BibTeX when Europe PMC has them, so you can pick later.

Europe PMC returns two journal-name fields – a full title and an NLM abbreviation. The tool uses the full title because that’s what most bibstyles (APA, MLA, Chicago, IEEE) expect to receive and then either render full or abbreviate themselves using their own rules. Bibstyles that want NLM-style abbreviations (Vancouver and most clinical journals) have package options to abbreviate from the full name. If you specifically need the NLM abbreviation in the source .bib, search-and-replace the journal field after copying.

Three common reasons: a typo in the PMID (PMIDs are 1 to 8 digits, no letters), a paper that was retracted and removed from MEDLINE (rare), or an article that exists in NCBI’s PubMed but hasn’t yet propagated to Europe PMC’s mirror. Each failure shows up inline with the original PMID and the reason. If Europe PMC reports “not-found” but you know the paper exists, try the same lookup via DOI on /doi-to-bibtex/. If a resolved entry compiles but throws a downstream error in your manuscript, our Build Doctor handles the diagnosis.

Preserved as UTF-8 in both dialects – there’s no LaTeX accent-macro equivalent for non-Latin scripts. They render correctly under biber with a Unicode-aware biblatex style, or under pdfLaTeX with usepackage[utf8]{inputenc} and a font that supports the script (use XeLaTeX or LuaLaTeX for the cleanest path). If your bibstyle isn’t Unicode-aware, those names will appear as missing glyphs in the compiled PDF regardless of the converter.

Free, no signup. PMIDs go directly from your browser to Europe PMC’s REST API – they never touch our server. The output exists in your browser session and disappears when you close the tab.

The clean BibTeX output is the right starting point, but publisher templates each enforce specific field orderings, citation styles, and field combinations – Elsevier’s elsarticle, BMC’s bmc-mathphys, Springer Nature’s sn-jnl. The tool doesn’t enforce per-publisher conformance. See our Elsevier and Springer formatting services for end-to-end submission conformance, or email help.thelatexlab@gmail.com with the target journal for a quote.

You Have a Word or PDF File. We'll Convert it to Submission-Ready LaTeX.

We convert Word documents and PDFs into clean, submission-ready LaTeX for IEEE, Springer, Elsevier, ACM, arXiv, and more. Equations, tables, citations, and journal templates included. Compiled and tested in Overleaf before delivery.

Reviewed by a real LaTeX specialist.