Importing and Exporting


Jump to: navigation, search


Can I import my existing references into CiteULike?

Yes. Create a BibTeX or RIS file containing your references and then import it into CiteULike by using the "Import" link in the "My CiteUlike" menu.

I posted a paper from my homepage, but it doesn't appear on the front page of CiteULike. What's wrong?

Nothing's wrong. If you posted a paper from a site that's not on the "supported" list, then it will appear in your library, but it won't be widely publicised on CiteULike (so it won't appear on the front page, and it won't appear in the "all papers for tag..." links). Of course, if you post a paper from one of our recognised journals, we know that the paper has already been peer-reviewed, and it will go on the front page.

This is hopefully an effective measure against both spam, and lunatics trying to peddle their crazy ideas about the origins of the universe.

The spam issue is obvious. If spammers worked out how to post links to the sites they're trying to advertise (and spammers are sometimes quite clever, if a little evil) and clutter up CiteULike with this junk then they'd obviously do it, and this site would be a worse place for it. Also, because we allow search engines onto CiteULike, the spammers might think that having lots of links to their site might boost their rankings in the Google search results. Actually, that wouldn't work because of a feature on Google... but they might not know that and try it anyway.

The "lunatics" problem is a bit more subtle. We regularly get emails from people who have posted articles (which clearly haven't been peer-reviewed) wanting to know what they can do to promote their work on CiteULike. The short answer is: nothing. These papers are generally either fairly curious ideas about the Big Bang, or results of less-than-scrupulous clinical trials of a drug (which is presumably manufactured by a company the user has shares in). CiteULike is not a place to try to push your own work to a larger audience (although, of course, you're welcome to submit your own articles and, if they're any good, others will bookmark them). This policy has the unfortunate side effect that some genuinely good articles might not be as visible as they might be. Hopefully this should slowly become less of an issue as as the number of supported journals increases.

Importing vs. Posting

Trusted versus untrusted data

At CiteULike we make a clear distinction between trusted and untrusted data. Trusted data has been taken by CiteULike directly from "approved" sites, most of which are peer-reviewed journal sites (listed here). All other data is treated as untrusted.

Why "untrusted"?

Our site is targetted by lots of spammers. By default, anything posted to CiteULike we can't authenticate is treated as potential spam.

A trusted article is extra special

Trusted articles:

  • appear on the front page
  • are returned through the main search
  • are used to determine neighbours.

Once you view a user's own pages, you can see all their articles, trusted and untrusted.

I've imported at bunch of articles into CiteULike. They come from trusted sites, but they seem to be untrusted.

When an article is posted to CiteULike, the URL can be in many different forms but which eventually points to the same page. For example:

  • Extra crud in the URL, such as a "session ID"
  • Web site aliases
  • Meta-sites which hold links to other publishers (such as the DOI system).

When the article is posted, we typically have to download the link and examine it to extract the specific metadata which will enable us to determine a unique link to the article. Once we have a unique link, we can then match up all users with the same article in their libraries.

However, when articles are imported (uploaded), it's impractical to process each one. One reason is that it would be impossibly slow for uploads of more than a few articles. Another is that we might get blacklisted by journals for overloading them with our requests.

So even imported articles from trusted journals are untrusted? Surely you can do better?

Yes we can. We're working on an "upload tool" which may mitigate many of the problems mentioned above.

How do I export my library (or sections of it) into a bibliography?

You can use the "Export" link in the "MyCiteULike" menu or the File:Export-icon.png‎ icon to export your library in either BibTeX or RIS format. You can then use BibTeX or EndNote (or whichever reference manager you prefer) to build it in to your bibliography.

You can also export your library as a PDF, RTF, or plain text, to include in a word processor, for example.

Another quick option to get a single citation into a word processor is to highlight the formatted citation, then "drag-and-drop" it directly into your word processor.

To export a section of your library tagged with a common tag, click on the tag name and then follow the steps above.

How do I export other libraries (or sub-libraries) into a bibliography?

You can use the File:Export-icon.png‎ icon and follow the steps above.

Note: There is a limit of 1000 articles per export when the export is for global libraries. These are general libraries and tag-libraries across CiteULike not associated with any user. They are the ones with URL addresses of the type: 

How can I access my CiteULike account from BibDesk?

Here's a blog entry from user dartar

How can I access my CiteULike account from Mendeley?

You can set up a live sync from CiteULike to Mendeley: any changes you make to your CiteULike library should appear in a special folder (on Mendeley Web, which can then be synced back to Mendeley Desktop). See the Mendeley help page for details on setting this up.

If there are any problems, please contact Mendeley Support in the first instance.

How can I access my library from application X?

You can get at the metadata in several ways from any HTTP client. Apart from the COinS record, we advise against parsing the raw HTML as the format may change without notice.

N.B. If you need to access any private articles/notes/groups, the application will need to "logon" first with your username and password. You can do this by emulating the logon form or contact us for a way to do this with a JSON API call.

Please contact us if you have an application you would like to integrate with CiteULike.

See also:


Your library is accessable through BibTeX and RIS downloads, which can be scripted. There's no magic, just emulate a browser using a HTTP library of your choice.

RDFa Header

Articles contain RDFa metadata in the <meta> headers.


With each article (including "list" views) is a COinS record, which can be parsed to get the metadata.

Below is a typical record:

<span class="Z3988" title="ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rfr_id=info:sid/


Your library is also available through RSS Feeds which contain all of the information to make a formatted citation.


Articles and article lists can be accessed as JSON(P) by prefixing the URL path with "/json", e.g.,

By default this returns the 50 most recent articles (unless overridden by sorting options). You can change the number of articles per call with per_page=N (N<250), and "page" through by setting page=1,2,..... You can set an optional JSONP callback function with callback.

Note JSON output should be considered experimental for now. We don't know what data you want yet and how you want to use it. Let us know so we can firm up the spec.

How can I export articles to Delicious?

You can export the current view to Delicious whenever you see the File:Export-icon.png button.

Save the file, and import into Delicious ("Settings" -> "Bookmarks" -> "Import..."). In general, we recommend importing into Delicious as "public" - any private articles will still be private.

Because of privacy and size restrictions we've had to make a few compromises:

  • Abstracts and notes are exported, but the former is truncated if the total length goes beyond Delicious's 1000 character limit.
  • Private notes are only exported if they're within private articles.
  • Tags are exported too, but not private (*xxxx) or for:xxxx tags.

Scripting CiteULike

In the following examples we show how to download the BibTeX file for a given library using wget - you can use similar URLs for exports of RIS, PDF, etc., and similar techniques can be used with other tools (curl, libwww-perl, etc)

To see private articles (or private notes) in your library, or to see articles in a private group, you must log in first:

wget -O /dev/null --keep-session-cookies  --save-cookies cookies.txt --post-data="username=xxxx&password=yyyy&perm=1"

Now that the login cookies are saved, you can access private data with, for example:

wget -O export.bib --load-cookies cookies.txt

The cookie doesn't expire so one only needs to log-in once (unless the password changes or - less likely - we change the cookie format). You could even copy the cookie ("login") from your browser.

NOTE: please be careful with automated scripting. Excessive load on our servers may result in automatic blocking of your account or IP address. 1000 requests per hour should keep you below the radar. (Note that some requests involve HTTP redirects - such calls count twice.)

Identifying your script

If you access CiteULike via an automated process, you MUST provide a means to identify yourself via the User-Agent string.

Please use "<username>/<email> <application>" e.g., "fred/ myscraper/1.0". Any scripting of the site without a means to identify you may result in a block.

Personal tools
MediaWiki Appliance - Powered by TurnKey Linux