Fueled by the growing popularity of proof-of-stake blockchains, there has been increasing interest and progress in permissioned consensus protocols, which could provide a simpler alternative to existing protocols, such as Paxos and PBFT. In particular, the recently proposed Streamlet protocol provides a surprisingly simple and streamlined consensus approach, which crystallizes years of research in simplifying and improving classical consensus protocols. While the simplicity of Streamlet is a major accomplishment, the protocol lacks certain practical features, such as supporting a stable block proposer, and it makes strong assumptions, such as synchronized clocks and the implicit echoing of all messages. Most importantly, it requires sending $O(N^3)$ messages per block in a network of $N$ nodes, which poses a significant challenge to its application in larger networks. To address these limitations, we introduce Pipelet, a practical streamlined consensus protocol. Pipelet employs the same block-finalization rule as Streamlet, but attains state-of-the-art performance in terms of communication complexity and provides features that are crucial for practical applications, such as clock synchronization and stable block proposers. At the same time, Pipelet retains the simplicity of Streamlet, which presents significant practical advantages, such as ease of implementation and verification.