The veeeb semantic editor currently uses the opencalais API to extract entities from a written text. But opencalais is not the only API out there. In fact there are different entity extraction API’s and a comparison can be very usefull.  Every API has some features another one doesn’t.

I really liked Rob DiCiucci‘s article about the current different semantic services. He covers more than just entity extraction in his article, e.g. the API of freebase, which veeeb is using as well or zemanta, which is more a API for suggesting relevant content directly (interessting, though…)

Regarding entity extraction the best known free API services right now seem to be

• http://www.alchemyAPI.com
• http://www.opencalais.com
• http://www.evri.com

In order to evaluate the services it is usefull to have the results on one screen directly. Thats why we build an easy Flex GUI where you can just do that.  It provides 2 lists of text – english and german. Doubleclicking an entry will execute the semantic services with the text of random news articles I found, but you can also type or copy&paste your own text.

Try it here:

entity extraction comparison tool

Remark: opencalais does not work with german text right now. So we have to proxy the requests via Google Translate to get some results. So do not blame opencalais for disturbing results for german text. But anyway – it works pretty good

I will post about the results of our comparison soon.

If you have any suggestions – let me know.

Happy extracting!

Christoph Diefenthal