BibTeX as an export format

Last reviewed on April 24, 2026

BibTeX is the oldest of the formats used here and still the default in LaTeX workflows. A .bib file is a plain-text database of bibliographic entries. Each entry has a type, a citation key, and a set of fields. A tool like biber or bibtex reads the file, matches citations from the main document against the keys, and produces a formatted bibliography.

Anatomy of an entry

@article{smith2023something,
  title   = {Something about pdf extraction},
  author  = {Smith, Jane and Doe, John},
  journal = {Journal of Document Engineering},
  year    = {2023},
  volume  = {12},
  number  = {3},
  pages   = {101--120},
  doi     = {10.1234/jde.2023.0012}
}

The @article announces the entry type. smith2023something is the citation key used from the main text (\cite{smith2023something}). The fields that follow are key-value pairs, with values wrapped in braces or double quotes.

Entry types

Original BibTeX defines a fixed set of types. The commonly used ones:

biblatex extensions

biblatex, the modern LaTeX bibliography system, reads BibTeX files and adds richer entry types — @online, @software, @dataset, @report, @collection — plus more fields (date that accepts full ISO dates, eprint with an eprinttype, language fields, URL access dates, and so on). Files produced for biblatex are still valid BibTeX files; they only fail when fed to the older bibtex backend that does not know the new types and fields.

The extractor on this site produces output that is safe for both: it uses @article for extracted articles and puts any ArXiv or PubMed identifier into dedicated fields that biblatex can pick up.

Authors, names, and the "and" separator

BibTeX uses the word and, case-sensitive, to separate author names. Commas inside an author entry are taken to mean "family first, given second". These are equivalent:

author = {Jane Smith and John Doe}
author = {Smith, Jane and Doe, John}

For names with suffixes or particles (van der Waals, de la Cruz, John Smith Jr.), use the explicit three-comma form: {von Part, Jr, First}. Getting this right matters for sorting and for style-specific formatting; misformatted names produce strange-looking bibliographies.

Citation keys

Keys have to be unique within a .bib file. A common pattern is lastnameYearShortTitle — smith2023pdfextraction. Keys should contain only ASCII letters, digits, and a small punctuation set; spaces and most Unicode characters cause trouble.

When multiple files are merged — for example, when combining several project bibliographies — a key collision is a silent hazard: biber will warn, but the behaviour is not always what you expect. Pick a project-wide key scheme early.

Encoding and special characters

UTF-8 BibTeX files work with modern toolchains (XeLaTeX or LuaLaTeX plus biblatex and biber). If your project is still on the classic pdflatex plus bibtex pair, you may need to escape accented characters using LaTeX commands: {\'e} for é, {\"u} for ü, and so on. Some math symbols inside titles need their braces carefully placed so that the title-case conversion in the style does not lowercase them.

Typical gotchas

When BibTeX is not the right choice

If the downstream tool is not LaTeX — for example, a web-based editor that uses CSL for formatting — exporting CSL-JSON directly avoids a round-trip through BibTeX's type system. See CSL-JSON as an interchange format. If the downstream tool is EndNote or a publisher's submission portal, RIS is usually a smoother fit; see RIS.

Related