Cuba-Memorys MCP Server

Example Use Cases

• 1

  Persistent long-term memory with adaptive decay and neuroscience-inspired learning.

• 2

  Knowledge graph-based reasoning and anti-hallucination verification.

• 3

  Error memory to avoid repeating past mistakes and session-aware context tracking.

Description

Cuba-Memorys is a Rust-based MCP server that enables AI coding assistants to retain long-term memory using advanced algorithms like FSRS-6 adaptive decay, Hebbian and BCM metaplasticity, and dual-strength memory models. It features a knowledge graph with typed relations, multi-signal fusion search, error memory to prevent repeated mistakes, and autonomous memory consolidation. The server supports semantic embeddings optionally via ONNX and offers robust anti-hallucination verification to improve AI reliability.