There are many tools around for annotating language corpora, but they tend to be good for one specific thing and they all use different underlying data formats. This makes it hard to mark up data for a range of annotations - disfluency and dialogue acts and named entities and syntax, say - and then get at the annotations as one coherent, searchable database. It also makes it hard to represent the true structure of the complete set of annotations. These problems are particularly pressing for multimodal research because fewer people have thought about how to combine video annotations for things like gesture with linguistic annotation, but they also apply to audio-only corpora and even textual markup. The open-source NITE XML Toolkit is designed to overcome these problems.

At the heart of NITE there is a data model that expresses how all of the annotations for a corpus relate to each other. NXT does not impose any particular linguistic theory and any particular markup structure. Instead, users define their annotations in a "metadata" file that expresses their contents and how they relate to each other in terms of the graph structure for the corpus annotations overall. The relationships that can be defined in the data model draw annotations together into a set of intersecting trees, but also allow arbitrary links between annotations over the top of this structure, giving a representation that is highly expressive, easier to process than arbitrary graphs, and structured in a way that helps data users. NXT's other core component is a query language designed specifically for working with data conforming to this data model. Together, the data model and query language allow annotations to be treated as one coherent set containing both structural and timing information.

Using the data model and query language, NXT provides: