Persistent memory for AI agents — 12 tools with Cuban soul.
Cuba-Memorys gives AI coding assistants long-term memory with a knowledge graph, Hebbian learning, GraphRAG enrichment, and anti-hallucination grounding.
What makes it different:
Knowledge graph — Entities, observations, and typed relations that persist across sessions Hebbian learning — Memories strengthen with use (Oja's rule) and fade adaptively (FSRS spaced repetition) Anti-hallucination — Verify claims against stored knowledge with graduated confidence scoring (verified/partial/weak/unknown) 4-signal RRF fusion search — TF-IDF + full-text + trigrams + optional pgvector HNSW GraphRAG — Top results enriched with degree-1 graph neighbors for topological context REM Sleep — Autonomous background consolidation after 15min idle Error memory — Never repeat the same mistake twice with anti-repetition guard Graph analytics — Personalized PageRank, Louvain communities, betweenness centrality, Shannon entropy Built on peer-reviewed math: Wozniak/Ye (FSRS), Oja (Hebbian), Cormack (RRF), Brin & Page (PageRank), Collins & Loftus (spreading activation), Malkov & Yashunin (HNSW).
Zero manual setup. Auto-provisions its own PostgreSQL via Docker. 12 tools, all named after Cuban culture.
Server Config
{
"mcpServers": {
"cuba-memorys": {
"command": "python",
"args": [
"-m",
"cuba_memorys"
],
"env": {
"DATABASE_URL": "postgresql://cuba:memorys2026@127.0.0.1:5488/brain"
}
}
}
}