cuba-memorys

Overview

Cuba-Memorys is a comprehensive MCP server that implements advanced memory features for AI agents, including a knowledge graph with typed relations, Hebbian learning, and exponential importance decay. The server offers 23 tools for memory management, hybrid search capabilities using RRF fusion, contradiction detection, and prospective memory triggers. It features a PostgreSQL backend with automatic Docker provisioning, making it easy to set up and use in various environments. The implementation demonstrates strong technical rigor with 97 tests, zero clippy warnings, and regular updates with new features like project scoping and LLM-judged contradiction resolution.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• cuba_alma

  CRUD entities with Hebbian boost + access tracking. Fires prospective triggers on access.

  Try:Add a new concept entity for blockchainTry:Update the person entity for John DoeTry:List all technology entities
• cuba_cronica

  Observations with semantic dedup, auto-tagging, session provenance, and contextual embedding.

  Try:Add an observation about today's meetingTry:List all observations from last weekTry:Find observations related to machine learning
• cuba_puente

  Typed relations. Traverse walks the graph. Infer discovers transitive paths. Predict suggests missing relations.

  Try:Create a relation between concept X and technology YTry:Traverse relations from entity ATry:Predict missing relations for entity B
• cuba_ingesta

  Bulk knowledge ingestion: arrays of observations or long text with auto-classification.

  Try:Ingest this research paper about AITry:Import these observations from a CSV fileTry:Process this long document with auto-tagging
• cuba_faro

  RRF fusion search with entropy routing, temporal filters, tag filters, and score breakdown.

  Try:Search for memories about project XTry:Find observations with tags Y and ZTry:Search for memories before date 2023-01-01
• cuba_sueño

  REM sleep consolidation: stratified decay + PageRank + auto-prune + auto-merge + episode decay.

  Try:Run consolidation on all memoriesTry:Consolidate episodic memories older than 30 daysTry:Auto-prune low-importance memories
• cuba_zafra

  Decay memories with testing effect: halflife scaled by access count frequency.

  Try:Apply decay to all observationsTry:Decay memories with low access frequencyTry:Update importance scores based on decay
• cuba_proyecto

  Isolate memories per project. List projects and manage project-scoped memories.

  Try:Start a new project called 'Apollo'Try:List all projectsTry:Switch to project 'Gemini'
• cuba_juez

  Escalate ambiguous-similarity observation pairs to LLM judge for contradiction resolution.

  Try:Judge potential contradictions in these observationsTry:Resolve ambiguous similarity score pairsTry:Cache verdict for future reference
• cuba_calibrar

  Bayesian calibration of source credibility with Beta(α,β) updates.

  Try:Update source credibility for data source XTry:Calibrate trust score for information from YTry:Track source reliability over time
• cuba_sync

  Export memories in git-friendly format or import idempotently.

  Try:Export memories for project XTry:Import memories from backup fileTry:Sync memories to external system
• cuba_faro

  Search with BM25 hybrid 3-way fusion, MMR diversification, and OOD abstention.

  Try:Search with BM25 and vector fusionTry:Diversify search results to avoid redundancyTry:Abstain from out-of-distribution matches

Comparable tools

Installation

pip install cuba-memorys

{
  "mcpServers": {
    "cuba-memorys": {
      "command": "cuba-memorys"
    }
  }
}

{
  "mcpServers": {
    "cuba-memorys": {
      "command": "cuba-memorys",
      "env": {
        "DATABASE_URL": "postgresql://user:pass@localhost:5432/brain"
      }
    }
  }
}

FAQ

What makes Cuba-Memorys different from basic memory MCP servers?

Cuba-Memorys implements advanced features like knowledge graphs with typed relations, Hebbian learning, hybrid RRF fusion search, graph intelligence analytics, contradiction detection, and neuroscience-inspired memory consolidation algorithms that are not found in basic memory implementations.

How does the server handle data persistence?

The server uses PostgreSQL as its backend with pgvector for vector storage. It can automatically provision a Docker container with PostgreSQL on first run, or connect to an existing PostgreSQL instance. Data is maintained across sessions with various decay mechanisms and consolidation processes.

Compare cuba-memorys with