The covid-19 pandemic that emerged in early 2020 has affected our activities with various limitations, for this reason the government has implemented several protocols starting from using masks, maintaining cleanliness, and social distancing. It is also undeniable that humans are social creatures, in this context boredom is the main enemy. There are some activities to be able relieve stress such as reading, listening to music, watching, or playing a game. TCG (Trading Card Game) is an artificial game that build in with such various interesting themes. Card games are generally played with other people, but what would happen if the card game was played in the covid-19 pandemic. Of course, the biggest obstacle we will be facing are distance and time. Konami as a card game manufacturer and developer has a brilliant idea by implementing Remote Duel where tournaments and other official matches can be held virtually. Cardfight!Vanguard is no exception, which has a gameplay that really depends on the interaction between players. Remote play requires players to use a camera that will take pictures of the card field while playing. This study uses the Waterfall Model which will be taken from dataset preparation, train the data, perform the Annotate process, and carry out a web-based implementation using the flask framework so that it can be used and tested for its functionality using one of tensorflow's product technologies, Roboflow, which is an application that specifically designed to be able to assist in the process of creating and recognizing objects. The results obtained by using flask as a web base can be seen to perform card object recognition properly so that it can display data from the detected cards.