Omega Memory

Persistent memory for AI coding agents

python

27

GitHub Stars

python

Language

6 months ago

First Indexed

4 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "omega-memory-omega-memory": {
      "command": "python3",
      "args": [
        "-m",
        "omega.server.mcp_server"
      ]
    }
  }
}

You run the OMEGA MCP server to give AI coding agents long-term, local memory across sessions. It captures decisions, lessons, and context, stores them securely, and surfaces relevant memories automatically as you work, enabling cross-session learning without relying on remote services.

How to use

You run the MCP server locally and connect your MCP client to it. If you have a remote client like Claude Code, you can register OMEGA as an MCP server and start storing memories that persist between sessions. Use the provided CLI to set up the MCP server and verify that it is healthy.

How to install

Prerequisites: Python 3.11 or newer, a working network for initial setup, and Git if you are installing from source.

git clone https://github.com/omega-memory/omega-memory.git
cd core
pip install -e ".[dev]"
omega setup

Configuration, security, and usage notes

Set up the MCP server so your MCP client can connect and begin saving memories. You can register the server via the command line interface of your MCP client when prompted to add a new server. The server runs locally and stores memories in an SQLite database with optional encryption.

Troubleshooting and tips

If you encounter issues during setup or operation, run health checks and verify the MCP registration. For example, you can verify the installation health with the included doctor command and check that the MCP server is registered for your client.

Architecture and CLI overview

OMEGA runs as an MCP server that exposes 12 memory tools for storing decisions, lessons, errors, preferences, and more. It uses a local SQLite database to store memories, embeddings, and edges for graph-based context. The server is designed to work entirely locally, without cloud dependencies.

Key CLI commands include setup, doctor, status, query, store, timeline, activity, stats, consolidate, compact, backup, validate, logs, and migrate-db. Hooks trigger automatic memory capture and timely surfacing of relevant memories during work.

Remote / SSH setup tips

If you need to run the MCP server on a remote machine, install omega-memory there, configure it with omega setup, and use a remote session to access the memory graph. The server’s memory persists across SSH reconnects, so you can resume work exactly where you left off.

Available tools

omega_store

Store typed memory items such as decisions, lessons, errors, preferences, and summaries.

omega_query

Semantic or phrase search with tag filters and contextual re-ranking.

omega_lessons

Cross-session lessons ranked by access counts for quick recall.

omega_welcome

Session briefing with recent memories and user profile.

omega_profile

Read or update the user profile.

omega_checkpoint

Save task state for cross-session continuity.

omega_resume_task

Resume a previously checkpointed task.

omega_similar

Find memories similar to a given memory.

omega_traverse

Walk the memory graph with typed edges.

omega_compact

Cluster and summarize related memories.

omega_consolidate

Prune stale memories and clean edges.

omega_timeline

Group memories by day for timeline view.

omega_remind

Set time-based reminders for tasks.

omega_feedback

Rate surfaced memories as helpful or not.