018 GitHub stars
02Token Management Policies: Includes active context ceilings and compaction triggers to optimize token usage for AI agents.
03Layered Memory Architecture: Integrates markdown for human truth, `qmd` for document retrieval, `Zvec` for semantic artifacts, and a metadata store for lifecycle state.
04Cloud-Assisted, Local-First Operation: Designed for hybrid environments, prioritizing cloud services with local fallback for resilience.
05Structured Reasoning & Session Persistence: Supports a First Principle thinking framework with durable session and project memory.
06Stable MCP Interface: Provides a consistent interface for agent-memory system interaction, including context retrieval, memory promotion, and session summarization.