MCPLow riskUnclaimed

memory-industry

MemoryIndustry persistent memory MCP server: BM25+MMR retrieval and audit.

LeandroPG19leandropg19/memory-industry★ 29Updated Oct 8, 2026

server.json

{
  "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
  "name": "io.github.LeandroPG19/memory-industry",
  "description": "MemoryIndustry persistent memory MCP server: BM25+MMR retrieval and audit.",
  "repository": {
    "url": "https://github.com/LeandroPG19/Memorys",
    "source": "github"
  },
  "version": "0.28.0",
  "packages": [
    {
      "registryType": "npm",
      "identifier": "memory-industry",
      "version": "0.28.0",
      "transport": {
        "type": "stdio"
      },
      "environmentVariables": [
        {
          "description": "PostgreSQL connection URL (e.g. postgresql://user:pass@localhost:5432/brain)",
          "isRequired": true,
          "format": "string",
          "name": "DATABASE_URL"
        },
        {
          "description": "Set to 'off' to disable v0.8 project scoping filters (admin/debug only).",
          "format": "string",
          "name": "CUBA_PROJECT_FILTER"
        },
        {
          "description": "Directory used by cuba_sync export/import (default ./.memory-industry/, falls back to ./.cuba-memorys/).",
          "format": "string",
          "name": "CUBA_SYNC_DIR"
        },
        {
          "description": "cuba_juez backend selector: nli | mcp_sampling | claude_cli | heuristic | auto (default auto).",
          "format": "string",
          "name": "CUBA_JUDGE"
        },
        {
          "description": "Subprocess CLI used by ClaudeCodeJudge (default 'claude').",
          "format": "string",
          "name": "CUBA_JUEZ_CLI"
        },
        {
          "description": "Model name passed to the judge backend (default 'claude-haiku-4-5').",
          "format": "string",
          "name": "CUBA_JUEZ_MODEL"
        },
        {
          "description": "Isolated entities the REM cycle scans for relations per run (default 5, 0 disables).",
          "format": "string",
          "name": "CUBA_REM_RELATION_BATCH"
        },
        {
          "description": "Directory holding the bge-reranker ONNX model; enables cross-encoder reranking (+93% nDCG on GPU).",
          "format": "string",
          "name": "CUBA_RERANKER_PATH"
        },
        {
          "description": "Pad reranker batches to a constant 512 tokens. Defaults on for GPU builds, off on CPU.",
          "format": "string",
          "name": "CUBA_RERANK_FIXED_SHAPE"
        }
      ]
    },
    {
      "registryType": "pypi",
      "identifier": "memory-industry",
      "version": "1.30.0",
      "transport": {
        "type": "stdio"
      },
      "environmentVariables": [
        {
          "description": "PostgreSQL connection URL (e.g. postgresql://user:pass@localhost:5432/brain)",
          "isRequired": true,
          "format": "string",
          "name": "DATABASE_URL"
        }
      ]
    },
    {
      "registryType": "npm",
      "identifier": "cuba-memorys",
      "version": "0.28.0",
      "transport": {
        "type": "stdio"
      }
    },
    {
      "registryType": "pypi",
      "identifier": "cuba-memorys",
      "version": "1.30.0",
      "transport": {
        "type": "stdio"
      }
    }
  ]
}

Permissions

DeclaredDetected
Runs code—nodepython
Installs—npm:memory-industry@0.28.0pypi:memory-industry@1.30.0npm:cuba-memorys@0.28.0pypi:cuba-memorys@1.30.0
Runs install scripts—npm:memory-industrynpm:cuba-memorys
NetworkNoneNone
Needs credentialsNoneNone
Outside the workspace—None
Agent tools—None

Checks

Low risk · Nothing worth a warning was found.

Not reviewed by a person · Checked by rules; the model review is not switched on yet.

Versions

  1. #20.28.0latestOct 8, 2026
  2. #10.27.0Oct 7, 2026