Show HN: Total Recall – write-gated memory for Claude Code (github.com)

A new plugin called Total Recall has been introduced for Claude Code, designed to enhance memory management by implementing a selective write-gated memory system. Unlike traditional memory tools that indiscriminately capture all inputs, Total Recall employs a strict five-point filter to ensure that only observations impacting future behavior, commitments, and stable facts are saved. This method first deposits information into a daily log before allowing users to promote entries to long-term memory, thus maintaining a lean and efficient context window. This innovation is significant for the AI/ML community as it addresses common issues related to memory bloat and hallucinations generated from irrelevant or outdated data. By fostering a more disciplined approach to information retention, Total Recall optimizes Claude Code's performance and reliability. It also prioritizes user control and correction—ensuring that corrections are immediately updated across all memory tiers without unintended overwrites. The plugin promotes a clearer, more efficient way to handle memory within complex AI systems, allowing for improved conversation accuracy and model behavior.