Wake Intelligence

概要

Wake Intelligence is a production-ready Model Context Protocol (MCP) server that implements a sophisticated temporal intelligence 'brain' for AI agents. It features a unique three-layer architecture: a Causality Engine for understanding past contexts and relationships, a Memory Manager for optimizing present relevance, and a Propagation Engine for predicting future context needs. This tool serves as a reference implementation for Semantic Intent as Single Source of Truth patterns, emphasizing observable properties and clear domain boundaries to create maintainable and AI-friendly codebases.

主な機能

• 4-tier memory classification (ACTIVE, RECENT, ARCHIVED, EXPIRED) with LRU tracking and automatic recalculation
• Causal chain tracking and reasoning reconstruction for 'why' contexts were created
• 1 GitHub stars
• 3-layer temporal intelligence system (Past: Causality, Present: Memory, Future: Propagation)
• Composite prediction scoring (temporal, causal, frequency) for proactive context pre-fetching
• Hexagonal Architecture with clear separation of concerns (Domain, Application, Infrastructure, Presentation)

ユースケース

• Enable AI agents to understand causal relationships and reconstruct past reasoning
• Optimize AI agent memory management by prioritizing relevant contexts and pruning expired ones
• Predict future context needs for AI agents to enable pre-fetching and caching strategies