isi2bibtex - convert ISI database files to BibTeX format


isi2bibtex [OPTIONS] inputfile [outputfile]

If no output file is specified, inputfile.bib is used.

Records are appended if the output file exists.


Isi2bibtex converts an ISI (Institute for Scientific Information) bibliographic database file to a BibTeX file for use in TeX and LaTeX documents. Both formats hold bibliographic data on scientific and other academic documents.

(In the UK, the ISI databases are commonly known as 'BIDS' or 'MIMAS WoS')

Another way to do the same job isi2bibtex does is with bp, which has the advantage of converting between many different bibliographic formats and character sets. If you don't want that, isi2bibtex understands BIDS standard format in addition to the others, and is stand-alone and so presumably easier to get working.


-h, --help display help and exit

-v, --version display version information and exit

-q, --quiet no informational output

-a, --abstract include abstract in output file

-c, --check make some checks on field contents (default)

-n, --nocheck don't make checks on field contents

Input databases

Although isi2bibtex was written for SCI (Science Citation Index), all the ISI databases should work (SCI, SSCI, A&HCI, ISTP). Isi2bibtex will probably make a bad job of editing the content of these other databases, and would have to be changed a bit (not difficult), but you may be lucky.

In the UK, probably most of the other databases on BIDS should work either straight away or with a small amount of modification of the script. BIDS Pascal for instance works with downloading format, and would work with a small amount of modification with standard format.

Input formats

If you use a web interface to ISI, isi2bibtex will only convert text output (whether emailed or saved directly), not saved web pages or other bibliography formats such as Procite or Reference Manager. Specifically, the formats that are understood are:

ISI generic output format version 1.0:

I presume this is the format used in most of the world. In the UK, this is output by MIMAS WoS 'save records' or 'email records'.

BIDS standard format:

(any of: Title only; Title, authors & journal; Full record excluding citations; Full record)

BIDS downloading format:

(any of: Title, authors & journal in downloading format; Full record in downloading format)

You can mix and match record formats in one file. Deleting fields and / or record numbers from records should be okay. Any fields may be present in any order. Don't change the indentation of records: isi2bibtex will ignore records if they're too different from the usual layout due to ambiguity of field labels and field contents.

Output fields

For ISI generic format (eg MIMAS WoS output) and BIDS downloading format, fields other than title (TI), author (AU), source ie journal (SO), page range (BP, EP), year (PY) and abstract (AB) are ignored.

For standard BIDS format, fields other than Title (TI:), author (AU:), journal (JN:) and abstract (AB:) are ignored (the JN: field contains the page numbers, volume, date, etc as well as the journal name).

At the moment isi2bibtex only outputs the more useful fields, but this may change in the future (when someone gets round to it). Which fields are output can be controlled with a configuration file (see below). Those that don't correspond to the standard BibTeX fields (such as abstract) won't be recognised by BibTeX by default, but they'll be there if you need to use them.

Output formatting

Output is tidied up as much as possible, but some editing is still required.

Output formatting defaults can be modified with a configuration file. /etc/isi2bibtexrc and ~/.isi2bibtexrc (or whatever you set in the @CONFIG variable in the script) are looked at in that order for configuration settings, with the latter overriding the former. See the example configuration file provided.

If they are switched on, things like journal title abbreviations and acronym capitalisations can be added and removed at the end of the script (very easy to do by looking at what's there already). Newer ISI entries have lower case as well as upper, and isi2bibtex always leaves the capitalization as-is for those records.


Remember to:


set the first line of this script to point to your copy of perl (eg. /usr/bin/perl or /usr/local/bin/perl)

make it executable (eg. chmod +x ./isi2bibtex)

put it somewhere your OS can find it (eg ~/bin or /usr/local/bin): you may need to change your PATH environment variable

Microsoft Windows

change the name of this script to (you'll have to use it as [OPTIONS] inputfile.txt outputfile.bib

rather than

isi2bibtex [OPTIONS] inputfile.txt outputfile.bib

because the windows DOS shell doesn't know about the #! line)

put it in your Perl bin directory, eg C:\Perl\bin (obviously if you don't have perl installed, you need to do that first: see below)

if that doesn't work check the PATH environment variable contains your perl bin directory, and as a last resort try

    perl C:\Perl\bin\ inputfile.txt

Other platforms

I have no idea, but it should work on any platform that runs Perl 5, perhaps with a few small modifications. Isi2bibtex has been tested on Windows (95 and NT) and Unix. Please send me any portability changes.

Email <> if it still doesn't work.


It only does articles, not books, proceedings etc, and won't notice if a record isn't an article.

It ignores some fields (mostly those that don't correspond to the standard BibTeX 'article' fields).

ISI access providers' output other than BIDS and MIMAS WOS haven't been tested. Send me an output file and I'll make it work with that format (for SCI).

Please tell me about any bugs you find, at <>


If you don't have Perl installed, it can be got (free) from . LaTeX, BibTeX and everything else TeX can be downloaded from .

bp converts between many bibliography database formats (including conversion from BIDS downloading and ISI generic formats to BibTeX), and many character sets.

Ben Bolker (author of the BIDS / MIMAS WOS specification for bp) has a page describing his modifications to bp:

Dana Jacobsen (author of bp) has a web page with lots of bibliography software information and details of his bp package:

Some other BibTeX utilities that may or may not be useful (these are just the ones I've got round to looking at):

bibclean checks and formats BibTeX databases

bibsort sorts BibTeX databases

bib2dvi converts BibTeX databases to DVI files (DVI files are output by LaTeX and are DeVice Independent typeset documents, a bit like postscript or pdf -- you can read them on most computer systems)

bibextract, citefind and citetags respectively extract BibTeX records from a BibTeX database, extract LaTeX references from a LaTeX document, and look up LaTeX references in a BibTeX database.

bibindex makes an index for fast lookup by biblook, if you have a huge database that needs it I suppose

BibTool is an all-singing all-dancing general purpose BibTeX management utility

bibview is a simple interactive searching utility for BibTeX files

findbib gets BibTeX records corresponding to references in a LaTeX file from a preprint server (don't know if it still works)

both refer2bibtex and r2bib convert refer files (whatever they are) to BibTeX files

tkbibtex is a graphical tool for editing and searching BibTeX databases

Text::BibTeX is a Perl module for doing things to BibTeX databases.


Copyright (C) 2000 Jonathan Swinton, Ben Bolker, Anthony Stone, John J. Lee

Isi2bibtex replaces and is derived from (as of 29 Jan 1998) and isi2bib 0.1. was based on a perl script written by Jonathan Swinton, and subsequently modified by Ben Bolker and Anthony Stone.

isi2bib 0.1 was written by John J. Lee <>

This script is covered by the GPL. See the script for copyright information.


/etc/isi2bibtexrc, ~/.isi2bibtexrc (or whatever you set in the @CONFIG variable) are looked at in that order for default configuration settings, with the latter overriding the former. See the example configuration file provided.