In the era of molecularly targeted therapy and immunotherapy, one common issue in phase I clinical trials is late-onset toxicity. A number of novel Bayesian adaptive designs have been proposed to address this issue, including model-based designs such as the time-to-event continual reassessment method (TITE-CRM), and model-assisted designs such as the time-to-event Bayesian optimal interval design (TITE-BOIN) and time-to-event keyboard (TITE-keyboard) design. In this talk, we will focus on TITE-BOIN and TITE-keyboard designs because of their simplicity and superior performance. We will describe the methodology of these designs and their software implementation using Shiny. We will use real-world examples to demonstrate the use of the online Shiny app to design phase I trials with late-onset toxicity. The Shiny app is freely available at www.trialdesign.org.