humanoid-memory

What problem does it solve?

Humanoid-memory provides a cognitive memory system for AI agents to consolidate, recall, and automatically manage long-term conversational context so agents remember important user facts and forget noise over time.

Core Features & Use Cases

• V-score consolidation combining coherence, significance, noise, and resonance to decide what to store.
• Tensor fusion that merges similar memories and HNSW-backed semantic search for fast vector retrieval.
• Ebbinghaus forgetting curve for exponential decay, LLM-based C/S/N evaluation, hooks to inject memories into requests, and per-agent persistent storage.
• Use Case: Keep track of user preferences across sessions, inject relevant prior interactions into prompts, and automatically purge decayed or low-value memories.

Quick Start

Use humanoid-memory to consolidate recent interactions and recall top relevant memories for the query "my preferences".