Cloud computing has become a cornerstone of modern technology, offering on-demand and scalable access to shared computational resources. As critical infrastructure increasingly migrates to cloud environments, concerns regarding data privacy and security have intensified. Users often need to trust service providers, despite risks like unauthorized access, insider threats, and data breaches. This study presents Git as a use-case of trust-dependent cloud services and investigates the potential of an end-to-end encrypted Git service on trust assumptions, security guarantees, and application performance. A prototype (E2Git) was designed to enforce end-to-end encryption at the application layer by encrypting repository contents and metadata on the client side, and automating key distribution across users and workstations, under the assumption of a fully covert adversary with complete access to the cloud infrastructure. The prototype’s security features and performance characteristics were evaluated in comparison with existing secure Git solutions. Although the E2Git prototype incurs elevated resource overhead during push operations, due to retransmission of a monolithically encrypted and compressed repository snapshot, it strengthens confidentiality and reduces server trust assumptions for repository contents and Git metadata. Notably, the E2Git prototype outperforms the standard Git client in terms of execution time during clone and fetch operations. These findings highlight the feasibility of trustless version control and the application of end-to-end encryption to secure sensitive data and enhance privacy in cloud environments.