Students at HTW Berlin University of Applied Sciences have created a new semi-automatic tagging website called akiwi in cooperation with Fotolia and Pixolution. Users simply upload an image, which is compared to a library of more than 22 million photos, and keywords are suggested based on similar images and input from the user.

Akiwi aims to be more accurate than totally automated tagging tools, but faster than manually adding keywords. After uploading an image by way of a drag-and-drop interface, users select a similar image from ten photos identified as visually similar. If none of the photos suggested are relevant, users can instead click on a keyword related to the photo, which refines the suggested images and keywords. This process continues until the user is satisfied that most suggested keywords are applicable to his or her photo.

From there, clicking 'Finalize' allows for a last selection of keywords to be made, which can then be copied to the user's clipboard with a click. In theory, everything is accomplished with a mouse and no typing is required, though keywords can be entered manually. Akiwi is free to use, and currently sits at version 1.1.