moai-cc-memory

Memory management in Claude Code involves separating working memory from long-term memory, using file-based persistence to store important context outside the active context window, and applying context budgeting to track token usage. This approach preserves knowledge across session boundaries without losing critical information.

Implement a memory capsule pattern that selectively consolidates and saves key context to files between sessions. This enables knowledge retention workflows where important information survives session resets, avoiding repeated reprocessing of the same data.

Context compression uses memory consolidation techniques to summarize and prioritize information, reducing token consumption while retaining essential knowledge. Combined with forgetting policies that deprioritize obsolete data, this maximizes useful context within limited budgets.

Yes. Automate memory management through file-based persistence systems that handle context seeding, selective memory consolidation, and cleanup on a configurable schedule. This removes manual intervention while maintaining prioritized access to active working memory.

Use active session context for immediate working memory during short interactions. Switch to file-based persistence for long-running sessions, multi-step workflows, or when context window limitations force tradeoffs between token budget and available information.

Yes. Memory persistence integrates with context seeding to retrieve and rehydrate consolidated long-term memory at session start, combining persistent knowledge retrieval with fresh working-memory setup for continued problem-solving.