In this talk, we will describe the development of the UCREL Semantic Analysis System (USAS) and how it can be used to automatically annotate text with semantic fields, applying contextually appropriate labels with a high degree of accuracy to not just words but multiword expressions. We will describe the initial manual development of the linguistic resources (semantic lexicons and associated context rules) and the semantic tagger (and disambiguation methods) for English in the 1990s. Then, we will illustrate how the USAS framework has expanded over the last 30 years to incorporate 12+ languages and describe methods to automatically bootstrap linguistic resources when adding new languages. Bringing the research up to date, we will present the latest version of pymusas (https://pypi.org/project/pymusas & https://ucrel.github.io/pymusas), a Python open source extensible framework, to allow researchers to annotate multiple languages and flexibly apply a growing set of methods for semantic annotation, along with a development roadmap and an example to illustrate how to add new languages such as Korean. Finally, we will motivate all this research and development effort by showcasing a variety of applications for semantic tagging, incorporating metaphor analysis via Wmatrix, qualitative survey analysis, political discourse analysis, and language profiling for applications in forensic linguistics.