This work presents the process of studying, prototyping and developing a Web application called Moodsic, which aims to perform the sentiment analysis of a personal report typed by the user and use the detected sentiment to search for a playlist consistent (or opposite) to it, while also giving users the possibility of saving those searches. The application was developed with the Django framework and by integrating services provided by Spotify and Hugging Face via API. The solution explores the connection between music and human sentiments, and makes advancements in the machine-learning field explicit. The product of this work was an application that is simple, but intuitive, functional and full of potential.