BibTeX vs RIS (and EndNote XML, CSL-JSON): which citation format to export
Short answer: choose BibTeX (.bib) if you write in LaTeX or Overleaf, RIS (.ris) if you are importing into EndNote, Zotero, Mendeley or a similar reference manager, and CSL-JSON if you write in Markdown with Pandoc or build software on top of citation data. EndNote XML is mainly useful for moving whole libraries in and out of EndNote. All of them carry the same core information (authors, title, container, year, volume, pages, DOI); they differ in syntax, in how precisely they describe fields, and in which tools read them.
Side-by-side comparison
| BibTeX (.bib) | RIS (.ris) | EndNote XML (.xml) | CSL-JSON (.json) | |
|---|---|---|---|---|
| Origin | LaTeX's BibTeX program (1985) | Research Information Systems (Reference Manager) | EndNote (Clarivate) | Citation Style Language project |
| Syntax | @article{key, title = {…}} | Two-letter tags, one per line: TI - … | XML elements | JSON objects and arrays |
| Record type | Entry type (@article, @book…) | TY code (JOUR, BOOK…) | ref-type | type (article-journal, book…) |
| Citation key | Yes, required | Optional (ID) | Record number | id, required |
| Authors | One field, names joined with and | One AU line per author | One element per author | Array of name objects (family, given) |
| Special characters | LaTeX escapes or UTF-8 (biblatex/biber) | Plain text, UTF-8 in modern tools | UTF-8 | UTF-8 |
| Main users | LaTeX, Overleaf, JabRef | EndNote, Zotero, Mendeley, databases | EndNote | Pandoc, Zotero, citeproc tools |
The same reference in each format
BibTeX:
@article{smith2023,
author = {Smith, Jane and Doe, John},
title = {Something about {PDF} extraction},
journal = {Journal of Document Engineering},
year = {2023},
volume = {12},
number = {3},
pages = {101--120},
doi = {10.1234/jde.2023.0012}
}RIS:
TY - JOUR
AU - Smith, Jane
AU - Doe, John
TI - Something about PDF extraction
JO - Journal of Document Engineering
PY - 2023
VL - 12
IS - 3
SP - 101
EP - 120
DO - 10.1234/jde.2023.0012
ER - CSL-JSON:
[{
"id": "smith2023",
"type": "article-journal",
"author": [{"family": "Smith", "given": "Jane"}, {"family": "Doe", "given": "John"}],
"title": "Something about PDF extraction",
"container-title": "Journal of Document Engineering",
"issued": {"date-parts": [[2023]]},
"volume": "12", "issue": "3", "page": "101-120",
"DOI": "10.1234/jde.2023.0012"
}]Practical differences that matter
- Field coverage. RIS and EndNote XML have dedicated fields for things like database name, accession number, access date and notes, which databases like to export. Classic BibTeX has a smaller field set; biblatex adds many more (
eprint,urldate,langid…). - Strictness. BibTeX breaks on unbalanced braces and needs care with special characters. RIS is forgiving but tags vary between vendors (
JO,JF,T2for journal titles), so imports can map fields differently. - Round trips. Every conversion can lose detail, such as reference types that have no exact equivalent. Keep one master library (in a reference manager) and export from it, rather than converting files from one format to another repeatedly.
- Multiple records. All four formats can hold many references in one file, so a whole search result or library can be exported at once.
Which one should I pick when a database asks?
- Importing into Zotero: RIS or BibTeX both work; RIS usually keeps more database fields. Zotero can export any of these formats later.
- Importing into EndNote: RIS (or EndNote's own format if offered).
- Importing into Mendeley: RIS or BibTeX.
- Writing in LaTeX/Overleaf: BibTeX.
- Writing in Markdown with Pandoc, or processing citations in code: CSL-JSON.
- Not sure: RIS. It is the most widely accepted import format among reference managers.
The in-browser extractor can export what it finds in a PDF as BibTeX, RIS or CSL-JSON, so you can pick whichever fits your tool.