Advertise
MCP.so
Sign In

Cuba Memroys MCP Server

@lENADRO1910

Repository

Basic information

Category

Memory & Knowledge

Transports

stdio

Author

lENADRO1910

Submitted by

Leandro Pérez

About Cuba Memroys MCP Server

Persistent memory MCP for AI agents — Knowledge graph + Hebbian learning + Anti-hallucination. 12 tools, 1 dependency, zero manual setup.

#mcp#ai-memory#knowledge-graph#graphrag#hebbian-learning#anti-hallucination#postgresql#rag#semantic-search#spaced-repetition#pagerank#python

Cuba Memroys MCP Server config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "cuba-memorys": {
      "command": "python",
      "args": [
        "-m",
        "cuba_memorys"
      ],
      "env": {
        "DATABASE_URL": "postgresql://cuba:memorys2026@127.0.0.1:5488/brain"
      }
    }
  }
}

Cuba Memroys MCP Server tools

Tool list not available yet

This listing does not currently have a tool list to display. This does not mean the server has no tools; check its documentation for supported capabilities.

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Cuba Memroys?

Cuba Memroys gives AI coding assistants persistent long-term memory using a knowledge graph with entities, observations, and typed relations, Hebbian learning, GraphRAG enrichment, and anti-hallucination grounding. It provides 12 tools named after Cuban culture and requires zero manual setup.

How to use Cuba Memroys?

It auto-provisions its own PostgreSQL database via Docker, so no manual configuration is needed. The server exposes 12 tools, each inspired by Cuban cultural references, that let AI agents store, retrieve, and reason over knowledge across sessions.

Key features of Cuba Memroys

  • Knowledge graph with entities, observations, and typed relations
  • Hebbian learning (Oja's rule) and FSRS spaced repetition
  • Anti-hallucination with graduated confidence scoring
  • 4-signal RRF fusion search (TF‑IDF, full-text, trigrams, pgvector HNSW)
  • GraphRAG enrichment with degree-1 neighbor context
  • REM Sleep consolidation after 15 minutes idle
  • Error memory with anti-repetition guard
  • Graph analytics: PageRank, Louvain, betweenness, Shannon entropy

Use cases of Cuba Memroys

  • AI coding assistants that remember project context across sessions
  • Agents that avoid repeating past mistakes using error memory
  • Long-term knowledge curation with autonomous consolidation
  • Graph-based retrieval-augmented generation with topological context
  • Continuous learning systems that strengthen frequently used knowledge

Frequently asked questions

1

How does anti-hallucination work?

The server verifies claims against stored knowledge using graduated confidence levels: verified, partial, weak, or unknown.

2

What runtime dependencies does it have?

It requires Docker to automatically provision a PostgreSQL database. The server is built on peer-reviewed math (FSRS, Oja's rule, RRF, PageRank, HNSW, etc.).

3

How does Hebbian learning operate?

Memories strengthen with use via Oja's rule and fade adaptively using FSRS spaced repetition, making important knowledge more retrievable over time.

4

Where is the memory data stored?

All persistent data, including the knowledge graph and embeddings, is stored in a PostgreSQL database that the server auto-provisions inside Docker.

5

Does Cuba Memroys work with any AI model?

The README describes it as giving AI coding assistants long-term memory, implying integration via the standard MCP interface, but does not specify model constraints.

Comments

More Memory & Knowledge MCP servers

Snipara

alopez3006

MCP connector for Snipara, the Project Brain for AI coding agents

1

ctxfile

ctxfile

Local-first MCP server that snapshots your project's working state into one context object. Open-core, privacy-first.

2

UmmahAPI

Yousuf Kazmi

Islamic source data for assistants: Quran, hadith, tafsir, duas, prayer times and qibla. Answers come from real sources with citations instead of recalled text, so verses and hadith are not hallucinated. 36,000+ hadiths

Context7 MCP - Up-to-date Docs For Any Cursor Prompt

upstash

Context7 Platform -- Up-to-date code documentation for LLMs and AI code editors

58.7K

M1K3

Round-Tower

Local AI agent for your Mac with a built-in MCP server: 18 tools for your notes and PDFs, a temporal memory graph, an on-device LLM and a neural voice. No account, no telemetry. No data centers required.

7
Advertise