omem

Overview

OMEM is a shared memory system that solves AI agent amnesia by providing persistent memory across sessions, devices, and agents. It features a three-tier Space architecture (Personal, Team, Organization) enabling knowledge flow with provenance tracking. The system implements sophisticated memory management including Weibull decay models, 11-stage hybrid retrieval pipelines, and intelligent reconciliation mechanisms. It can be self-hosted or used via the hosted ourmem.ai service.

Try asking AI

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

Comparable tools

Installation

{
  "mcpServers": {
    "omem": {
      "command": "npx",
      "args": ["@ourmem/mcp-server"],
      "env": {
        "OMEM_API_KEY": "your-api-key-here"
      }
    }
  }
}

FAQ

What makes OMEM different from other AI memory systems?

OMEM's unique three-tier Space architecture enables cross-boundary knowledge sharing with provenance tracking. It combines sophisticated memory management (Weibull decay) with high-quality retrieval (11-stage pipeline) and seamless agent-to-agent sharing capabilities.

How does OMEM handle memory conflicts?

OMEM implements a 7-decision reconciliation system: CREATE, MERGE, SUPERSEDE, SUPPORT, CONTEXTUALIZE, CONTRADICT, or SKIP. This allows memories to evolve intelligently over time, with the system automatically determining how to handle conflicting or overlapping information.

Compare omem with