Show HN: Dejavu – stop showing coding agents the same command output twice (github.com)

Dejavu has been introduced as a command-output memory layer designed to optimize interactions with coding agents like Claude Code, Codex, and Gemini CLI. By tracking previously executed commands, Dejavu returns only meaningful changes in output rather than repeating the entire output for commands that produce the same results. This innovation is significant for the AI/ML community as coding agents frequently rerun commands during tasks such as testing and logging, where much of the output remains unchanged. With an early benchmark showing a reduction of 52-55% in output across repeated sessions, Dejavu streamlines the communication flow, enabling agents to focus on crucial information. The tool operates without needing explicit prompts from the agents, maintaining exit codes and storing full output locally for later retrieval. Compatible with popular commands like npm, git, and docker logs, Dejavu's shims allow seamless integration, enhancing the efficiency of terminal-based workflows. By ensuring that each command is still executed fully while selectively reducing the displayed output, Dejavu not only conserves context but also helps prevent information overload in coding environments. This approach promotes a more efficient and user-friendly experience for developers relying on AI-assisted coding tools.