omega-memory

Overview

OMEGA provides persistent memory for AI coding agents that works locally-first without cloud dependencies. It stores decisions, preferences, and architectural choices across sessions, eliminating context loss and repeated mistakes. The system runs entirely on the user's machine using SQLite storage with ONNX embeddings, avoiding vendor lock-in and data privacy concerns. As an MCP server, it integrates seamlessly with Claude, GPT, Gemini, Cursor, Claw Code, and any MCP client.

Try asking AI

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

• omega_store

  Store typed memory (decision, lesson, error, summary)

  Try:Remember that the auth system uses JWT tokens, not session cookies
• omega_query

  Semantic search with tag filters and contextual re-ranking

  Try:What did I decide about authentication?
• omega_welcome

  Session briefing with recent memories and profile

  Try:Give me a summary of my recent work decisions
• omega_profile

  Read or update user profile

  Try:Update my profile to prefer TypeScript strict mode
• omega_delete_memory

  Delete a specific memory by ID

  Try:Remove memory about the authentication system decision
• omega_edit_memory

  Edit the content of a memory

  Try:Update the memory about JWT tokens to include expiration time
• omega_lessons

  Cross-session lessons ranked by access count

  Try:Show me the most frequently accessed lessons
• omega_timeline

  Memories grouped by day

  Try:What decisions did I make last week?
• omega_compact

  Cluster and summarize related memories

  Try:Summarize all memories about authentication
• omega_checkpoint

  Save task state for cross-session continuity

  Try:Checkpoint the current refactoring task
• omega_resume_task

  Resume a previously checkpointed task

  Try:Resume the refactoring task from yesterday
• omega_remind

  Set a time-based reminder

  Try:Remind me tomorrow to review the authentication implementation

Comparable tools

Installation

pip install omega-memory[server]    # Full install (memory + MCP server)
omega setup                         # Downloads model, registers MCP, installs hooks
omega doctor                        # Verify everything works

pip install omega-memory[server]
omega setup --client claude-desktop

FAQ

What makes OMEGA different from Anthropic's Memory Tool?

OMEGA works with any LLM/agent (not just Claude), stores data entirely locally without requiring API calls, and includes advanced features like semantic search and multi-agent coordination.

Can I use OMEGA without running it as an MCP server?

Yes, you can install it as a Python library only with `pip install omega-memory` or use hooks-only mode with `omega setup --hooks-only` to save resources.

Compare omega-memory with