HeuroLabs started the private beta program for their API featuring computer vision, speech processing and entity extraction at a very high level. Try the live demo and the documentation. The API is designed to be very easy, with the main building block being the concept of a pipeline. Content is ingested to a pipeline and a key is returned. The key can be used to poll for results.
Overview: <a href="https://heurolabs.atlassian.net/wiki/display/CogniDoc/Overview" rel="nofollow">https://heurolabs.atlassian.net/wiki/display/CogniDoc/Overvi...</a>
API Docs:
<a href="http://api.cognitio.heurolabs.com/assets/docs.html" rel="nofollow">http://api.cognitio.heurolabs.com/assets/docs.html</a>. We are looking for feedback from the community, especially people working with hardware that is sensor rich as well as mobile applications that deal with visual/audio signals.