GPS sensors have been widely used to track people's everyday life trajectories, generating massive trajectory datasets. The trajectory data typically contains sparse GPS points, and completing trajectories is often necessary. State-of-the-art methods [3, 4] essentially complete the entire route by using a single metric, e.g., either the shortest distance or the fastest driving/walking time. Unfortunately, using a single metric may not always work in real life due to the diversity of mobility patterns. In this demo abstract, we propose a frequent pattern (FP)-based trajectory completion approach, and demonstrate a system prototype to showcase the advantages of our approach over four previous works, in terms of accuracy and running time.