MCP Catalogs
Homememorix screenshot

memorix

by AVIDS2·449·Score 53

Open-source cross-agent memory layer for coding agents via MCP with local-first storage.

developer-toolsai-llmknowledge-graph
39
Forks
8
Open issues
this month
Last commit
2d ago
Indexed

Overview

Memorix is a local-first memory control plane for coding agents that maintains project context, reasoning history, and Git-derived facts in one persistent location. It provides three-layer memory (observation, reasoning, and Git memory), source-aware retrieval, and a memory quality pipeline with deduplication and retention policies. The system supports multiple IDEs and coding agents like Cursor, Claude Code, Codex, Windsurf, and GitHub Copilot through MCP integration, enabling continuity across sessions and environments.

Try asking AI

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

you:Cross-session and cross-IDE project memory for coding workflows
you:Git-aware memory tracking and retrieval in development projects
you:Multi-agent coordination with task management and inter-agent communication
you:What makes Memorix different from standard IDE memory?
you:How does Memorix handle privacy and data storage?

When to choose this

Choose Memorix when you need persistent memory across multiple IDEs and coding agents, especially if you work with multiple AI coding tools and want them to share project context and reasoning.

When NOT to choose this

Don't choose Memorix if you only use one IDE and coding agent, prefer cloud-based solutions over local-first storage, or need a simpler memory system without the complexity of multi-agent orchestration.

Tools this server exposes

12 tools extracted from the README
  • memorix_store

    Store a memory in the memorix system with text, title, entity, and type

  • memorix_search

    Search through stored memories using a query

  • memorix_detail

    Retrieve detailed information about a specific memory

  • memorix_resolve

    Resolve or update a memory in the system

  • memorix_session_start

    Start a new session with memorix, optionally joining the agent team

  • memorix_task_list

    List available tasks in the agent team

  • memorix_team_status

    Check the current status of the agent team

  • memorix_orchestrate

    Run a multi-agent orchestration with a specific goal

  • memorix_sync_workspace

    Sync workspace configurations and rules

  • memorix_ingest_commit

    Ingest Git commit information into memorix

  • memorix_audit_project

    Audit the current project state and memory consistency

  • memorix_dashboard

    Launch the memorix dashboard UI

Note: Tool names inferred from CLI commands and MCP references in the documentation. The README doesn't provide explicit MCP tool definitions but references MCP tools throughout the text.

Comparable tools

semantic-memory-mcpdev-assistant-mcpcursor-memorygithub-copilot-memory

Installation

Install globally with npm:

npm install -g memorix

Initialize config:

memorix init

For MCP integration in Claude Desktop, add to config.json:

{
  "mcpServers": {
    "memorix": {
      "command": "memorix",
      "args": ["serve"]
    }
  }
}

Run as MCP server: memorix serve

FAQ

What makes Memorix different from standard IDE memory?
Memorix provides a persistent, cross-session memory layer that extends beyond single-threaded agent memory, with Git integration and multi-agent coordination capabilities.
How does Memorix handle privacy and data storage?
Memorix uses SQLite as its canonical store with Orama for search, keeping all data local with no cloud dependency by default.

Compare memorix with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.