Better Memory

Semantic memory, intelligent compression, and context management for AI agents.

What It Does

• Stores memories with real vector embeddings (local, no API calls)
• Semantic search via cosine similarity
• Auto-deduplicates on store (exact + semantic)
• Priority-based compression when approaching context limits
• Identity persistence across sessions
• Token-budget-aware memory retrieval
• Configurable context limits, thresholds, and summarizer

Quick Start

import { createContextGuardian } from 'context-guardian';