Since June 2012 LAF (stands for Linguistic Annotation Framework) has been accepted and ranked as ISO 24612:2012 standard providing an architecture for creating, multiple-annotating and managing linguistic resources. The framework specifies the glossary, the common mechanisms of the abstract data model and methods for descriptions of multiple linguistic data annotations. LAF defines the mechanisms for implementing constraints in an annotation scheme, but it does not have specific constraints for individual annotation schemes.

The abstract data model of LAF can be built by using various kinds of annotation formats. Different user-defined annotation formats can be mapped to the LAF abstract data model with the help of the XML serialization format, called the Graph Annotation Format (GrAF). GraF is a pivot format for mapping from an annotation scheme into another annotation scheme. In this way, users can choose their own annotation formats, transfer them into the pivot format for further interchange and/or processing, and eventually to transfer it back into the original format.

In the future, LAF should be able to represent most of linguistic information of language resources for different media types. The primary data and the annotation documents will be separated.