Memory

This is a persistent memory layer for AI assistants that goes beyond vector search. It gives you 25 tools including semantic search, identity tracking, contradiction detection, and knowledge graph traversal. The `velixar_context` tool synthesizes workspace briefings in one call, while `velixar_distill` extracts durable memories from sessions with automatic deduplication. It ships with 16 workflow prompts for orientation and conflict resolution, plus 5 live resources that inject recent and relevant memories into context without explicit tool calls. Memories are workspace-scoped (via git root, env var, or config file), and it includes circuit breakers and cache fallback for reliability. Reach for this when you need your assistant to remember preferences, track evolving beliefs, or maintain continuity across sessions without rebuilding context manually.