A voting system is responsible for collecting and recording voter intentions accurately to ensure the election outcome reflects the citizens’ will. Unfortunately, mistrust in voting systems has increased in the last decade. Although most mistrust is caused by false claims, some doubts are legitimate, and questioning the correctness of an electronic voting procedure is fair. A solution to mitigate these doubts is the use of End-to-End (E2E) verifiable voting systems that produce proof, which is verifiable by each voter, that the intention of voters is correctly captured, recorded, and used to compute the election outcome. Our goal in this work is to create an in-person E2E verifiable voting system that is well-suited for scenarios where each voting machine produces its own verifiable tally and a central authority aggregates these results to produce the final election outcome. One example of this scenario is the Brazilian Election. In hardware similar to the Brazilian voting machine, our scheme can create and cast votes in milliseconds, causing a negligible overhead to voters in the machine’s response time. Moreover, the most computationally heavy part of our protocol, the creation of Zero-Knowledge Proofs, requires approximately 6 min to be executed. However, this process is executed only once at the end of the voting period, and it does not require voters’ participation.