Paste a DOI (up to 50, resolved in parallel), it goes straight from your browser to CrossRef's API, comes back as bibliographic data, gets written to BibTeX. Nothing touches a server in between - exists in the browser session, gone when you close the tab.
Two output dialects: BibLaTeX keeps accents as raw UTF-8 (for biber), Legacy BibTeX rewrites them as macros like Kr{\"a}mer (for old pdfLaTeX bibtex setups, since raw UTF-8 either errors or sorts wrong there). Check your preamble - \usepackage{biblatex} + \addbibresource means BibLaTeX, \bibliographystyle + \bibliography means Legacy.
The actual fixes: titles aren't brace-protected by most converters, so under bibstyles that case-fold titles, "DNA repair in RNA viruses" becomes "Dna repair in rna viruses" - this wraps all-caps tokens in braces so that can't happen. CrossRef returns bare reserved characters like & in journal names, which is an active LaTeX character and throws "Misplaced alignment tab character" on compile - auto-escaped. CrossRef also mistags entry types more than expected (tags conference papers as journal-article), so there's a per-row dropdown to override @article before you copy.
Failed lookups show inline with the reason - usually a typo, a withdrawn paper, or a record not in CrossRef's index. If it's not in CrossRef but you know the PMID or arXiv ID, the PubMed and arXiv tools exist for exactly that fallback. Worth knowing: the BibTeX this produces is a clean starting point, not pre-conformed to IEEE/Elsevier/ACM submission templates - those enforce their own field ordering on top.
Link: thelatexlab.com/doi-to-bibtex/
SOCIAL SHARE CARD GENERATOR