Looking for properly formatted BibTeX citations for articles or books can be tedious. I used to use the functionality from Google Scholar, but found that for many cases the BibTeX files are poorly formatted, are missing information, or even contain wrong information.

For mathematics papers, http://www.ams.org/mathscinet/ appears to be a reasonable alternative. But it is manually maintained and unfortunately a paid service (many universities have access).

Are there other similar databases, possibly for other fields of science?

  • 1
    Kieran Healy maintains a database for some topics in sociology (kieranhealy.org/socbibs.html) and contributes to a database of philosophy papers and books (kjhealy.github.io/bib).
    – DG'
    Sep 6, 2013 at 17:02
  • 2
    not sure what you mean "manually maintained" for mathscinet; yes, the database entries are added manually, but the bibtex presentation is generated programmatically from the database. (i'm not sure how else what is close to a primary source can be created other than manually, unless you can guarantee that all participating publications/publishers adhere to exactly the same tagging conventions; if you can figure out how to make that happen, i'm sure math reviews would be delighted.) Sep 6, 2013 at 18:22
  • 3
    I think it's worth adding here that (1) this is borderline for on-topic (might fit equally on the Academic StackExchange site) (2) BibTeX data is one part of a bigger picture if you are creating reference database, and that is a serious business so free sources are limited (I would look to for example Web of Knowledge) (3) citations should be based on the primary literature, so there is an argument you should always get the data from the papers themselves.
    – Joseph Wright
    Sep 6, 2013 at 18:33
  • 1
    For computer science an excellent source is dblp.
    – Guido
    Sep 6, 2013 at 20:17
  • 2
    this is a bit late, but perhaps still useful. in addition to the paid access to the full mathscinet, data can be retrieved in bibtex format using the free service ams.org/mref Mar 24, 2016 at 20:18

4 Answers 4


For me, I do not use a single source, rather a combination of resources and techniques. Actually, there are a number of tools or resources you can use under the circumstances.

  • Documents searched through Cite Seer X will have a BibTeX entry at the right hand side of screen. Perhaps this is the most comprehensive one.

enter image description here

enter image description here

enter image description here

  • Entries in ACM Digital Library all has BibTeX entries. Please look at the right hand side of screen.

enter image description here

  • I am not sure which BibTeX entries in Google Scholar you are referring to. If you click on the actual publisher site (not the unreliable download link which may or may not appear at the right hand site), you will often find BibTeX entries in the publisher site. The point is also applicable in the previous case.
  • If you use menedely as your reference management tool, citing a paper from your personal collection is very easy.

enter image description here

Please see here how you can do this in mendeley. You can even generate a .bib file in one go, which is rather nice. Other reference management tools like Papers, EndNote, or Zotero should have similar functionalities, I am not very sure.


www.bibtexsearch.com has tens of millions of BibTeX records, covering mainly medicine, computer science, philosophy and finance.

  • 3
    When using this resource, be prepared to do a lot of extra manual editing of the entries it comes up with. A couple of casual searches I performed indicate that most (all?!) entries are given as being of type @article, even though many entries clearly should be of type @book, @techreport, @inbook, @incollection, etc.
    – Mico
    Apr 8, 2015 at 22:57

I know that I am resurrecting a century old question but I just become aware that Nelson H. F. Beebe wrote a tool bibclean which is a syntax checker for BibTeX files.

My favorite way of looking for BibTeX files is of course MathSciNet but it looks like somebody wrote a search engine just for BibTeX files


  • 3
    There's also betterbib. Oct 31, 2016 at 9:08
  • Note that betterbib is not open source. You will need to subscribe to Tex World! Mar 16, 2023 at 1:43

As of now, none exists.

E.g., if you search for "model checking" in http://www.bibtexsearch.com/, you would notice BibTeX entries for publications that lack certain information, such as the addresses of publishers.

Screenshot of a search on BibTeX Search for the term "model checking"

Prof. Jason Mars (https://midas.umich.edu/faculty-member/jason-mars/) had a script (a Python script, or a similar programming language) that could grab information from the ACM Digital Library (https://dl.acm.org/).

Unfortunately, I noticed that the ACM Digital Library can contain incorrect information for names (diacritics problem), titles (mathematical expression typesetting problem), publishers (diacritics problem), and addresses (diacritics problem) that have diacritics, diacritical marks, accents, or mathematical notations. Likewise, other digital/online libraries from other publishers have the same problem.

This is true, regardless of whether they provide access to the suggested BibTeX entry for each publication in the digital/online library.

I believe that cultural/ethnic bias also plays a role in complicating the ability of these digital libraries to track authors over time to determine who claims to be the (co-)author of a publication, due to people can use different versions of their names. E.g., using the initial (or first letter) of their first name, or initials for the first and middle names, or list all their middle names for a subset of their publications but not others. For people who come from populations with many people have the same first name and last name, but different middle names (if any, at all), this becomes complicated when people are not consistent in how they name themselves. People do change their names, not necessarily when they get married or change their gender identity, but also for other reasons. For people who live in English-speaking countries that have had to work with people who use computer technology, or typewriters back in the day, they may have used versions of their names without diacritics, but now do.

E.g., does "J. Smith" for a publication refers to John Smith, Jason Smith, or another Smith. What happens when people with the same tuple/set of first name and last name work in a research topic?

Fancy natural language processing, computer vision (for PDF only publications that lack text-only versions of the publications), and other artificial intelligence technologies, can help us with the aforementioned problems, including validating information in a given BibTeX entry with information that can be obtained from a valid and decent (quality of the images, PDF document, or physical copy of the publication/book) copy of the publication.

While we can all share our BibTeX databases, or set of BibTeX entries, do you trust that other people have done due diligence in validating information that they included in their BibTeX databases? After all, I just indicated the problems of publishers having problems validating/checking the information of BibTeX entries for their publications, including those associated with professional societies in computer science and computer engineering, or related field (such as electrical and computer engineering), such as ACM and IEEE.

However, start now, and over time you can accumulate tens of thousands of BibTeX entries in your BibTeX database. I have a BibTeX database with more than 20,000 BibTeX entries.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .