confaBOT explores the relationship between text and image by exploiting how images are currently dealt with on the internet. through the use of metadata [keywords + tags] our digital media and images are obsessively annotated with text in order to retrieve them in the future. this data is typically queried through the use of a search bar to find the particular content. confaBOT works in a similar way, however through the medium of an online chat where each line of conversation queries and displays the single most relevant image. the current version uses a rendition of ALICE chatbot – hoping that it would help keep the conversation flowing while providing surprising feedback and thus surprising imagery.

video demo: http://vimeo.com/31332817

features:
- chat for unlimited time with your neighborhood bot named confaBOT
- watch with surprise as images are presented for each line of chat you type
- click on any line to view a thumbnail of the chat image
- click a 2nd time to view the images original context
- email chat logs with links to each image to friends and family!

UPDATE: Ran out of API transaction credits, but the problem has been fixed and confaBOT should be working fine now! Sorry for the downtime...