FUTO Swipe is a new swipe typing model that aims to provide fast and accurate typing on mobile devices. The model is open-source and has been released under a permissive license, allowing the community to use and contribute to it. The model consists of three components: an encoder, a context language model, and a decoder, which work together to provide accurate predictions. The model has been trained on a dataset of 1 million swipes and has achieved a top-4 fail rate of around 4% on the test set.