[MCP Server] OMEGA - Persistent Memory for AI Agents

[singularityjason]

MCP Server Information

• Name: OMEGA
• GitHub: https://github.com/omega-memory/omega-memory
• Website: https://omegamax.co
• PyPI: https://pypi.org/project/omega-memory/
• License: Apache-2.0
• Logo: https://raw.githubusercontent.com/omega-memory/omega-memory/main/assets/omega-icon-400x400.png

Description

OMEGA is a local-first persistent memory system for AI agents. It stores decisions, preferences, lessons, and project context across sessions with semantic search, intelligent forgetting, and multi-agent coordination.

Key Features

• 25 MCP tools — memory, coordination, knowledge graph, reminders, checkpoints
• 95.4% on LongMemEval (ICLR 2025) — highest published score
• Zero cloud dependency — SQLite + ONNX embeddings, no API keys needed
• AES-256 encryption at rest
• Works with Claude Code, Cursor, Windsurf, LobeChat, and any MCP client

Installation

pip install omega-memory

MCP Server Config

{
  "mcpServers": {
    "omega": {
      "command": "python3",
      "args": ["-m", "omega", "serve"]
    }
  }
}

Manifest

{
  "author": "omega-memory",
  "homepage": "https://github.com/omega-memory/omega-memory",
  "identifier": "omega-memory",
  "locale": "en-US",
  "meta": {
    "avatar": "https://raw.githubusercontent.com/omega-memory/omega-memory/main/assets/omega-icon-400x400.png",
    "tags": ["memory", "mcp", "ai-agents", "persistent-memory", "local-first"],
    "title": "OMEGA - Persistent Memory for AI Agents",
    "description": "Local-first persistent memory with semantic search, AES-256 encryption, and 25 MCP tools. Zero cloud dependency."
  }
}

Links

• Docs: https://omegamax.co/docs
• Compare: https://omegamax.co/compare
• Smithery: https://smithery.ai/servers/omega-memory/core
• Existing listings: Smithery, Glama, mcpservers.org, PulseMCP, PyPI

cc @arvinxx @ONLY-yours @nekomeowww