A syntactically-guided sentiment analysis systems that are efficient, thanks to the breakthroughs in parsing speed obtained within the ERC Starting Grant FASTPARSE. Thanks to a new parsing paradigm based on sequence labeling and to distillation techniques, SALSA can improve the speed of state-of-the-art parsing by between 1 and 2 orders of magnitude. This means that it is now viable to implement systems that can syntactically parse sentences and use the obtained syntactic structures to perform accurate sentiment analysis, at speeds close to 1000 sentences per second on standard consumer hardware, thus democratizing accurate sentiment analysis technology and putting it within reach of smaller companies and institutions that cannot afford to deploy BERT-like models or slow parsers for large-scale processing.