<![CDATA[Low student numeracy skills require interactive learning innovations. Jarimatika is effective in aiding counting comprehension, but its conventional application lacks direct feedback and competitive elements. This study aims to design a web-based Jarimatika learning platform integrating Computer Vision technology and real-time network communication. The system development uses the Prototyping model with Black Box testing method. The system implements MediaPipe Hands on the client-side to detect 21 finger gesture landmarks automatically and convert them into Jarimatika values. To facilitate the competition mode, the WebSocket protocol is used to synchronize data between players in real-time, combined with a First-Come-First-Served (FCFS) algorithm to determine the winner. The test results show that the platform successfully detects finger formations accurately and maintains the stability of multiplayer synchronization with a latency of under 500 milliseconds. In conclusion, the integration of artificial intelligence and WebSocket has proven capable of producing an interactive, competitive, and adaptive Jarimatika learning media for users. ]]>
Copyrights © 2026
