At VRAIN, research in natural language processing (NLP) spans speech recognition and synthesis, machine translation, sentiment analysis, and text simplification. Work in this area addresses the challenges of multilingual processing and complex linguistic phenomena including irony detection and automatic summarisation.

Projects developed at VRAIN include TransLectures-UPV (TLP), a system for automatic multilingual transcription and subtitling of educational video content; SENTAT, for sentiment analysis in Spanish-language social media; and EN-IRONIC/ES-IRONIC, for irony detection. The TT-Streaming platform provides an RPC API for real-time transcription, translation, and dubbing of live audio broadcasts, complemented by neural network-based models for speech synthesis and recognition across multiple languages.