In this article we propose a local descriptor for an unconstrained handwritten word spotting task. The presented features are inspired by the SIFT keypoint descriptor, widely employed in computer vision and object recognition, but underexploited in the handwriting recognition field. In our approach, a sliding window moves from left to right over a word image. At each position, the window is subdivided into cells, and in each cell a histogram of orientations is accumulated. Experiments using two different word spotting systems – hidden Markov models and dynamic time warping – demonstrate a very significant improvement when using the proposed features with respect to the state-of-the-art ones.