Amem

Name: Amem
Author: amanasmuei

@amanasmuei

About Amem

The memory layer for AI coding tools. Local-first, semantic, 9 MCP tools with consolidation and project scoping. Works with Claude Code, Cursor, Windsurf & any MCP client.

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "amem": {
      "command": "npx",
      "args": [
        "-y",
        "@aman_asmuei/amem"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Amem?

Amem (Aman's Memory) is an MCP server that gives AI coding assistants persistent memory. It remembers user preferences, decisions, corrections, and project knowledge across conversations so the user never has to repeat themselves. All data stays on the local machine using SQLite and local embeddings, with no cloud or API keys required.

How to use Amem?

Install the package globally with npm install -g @aman_asmuei/amem, then configure it as an MCP server in your AI tool (Claude Code, Cursor, Windsurf, or any MCP client) by pointing to the amem command or npx @aman_asmuei/amem. Once connected, ask the AI to store or recall memories using built-in tools like memory_store and memory_recall.

Key features of Amem

Persistent memory for AI assistants via MCP.
Six memory types: corrections, decisions, patterns, preferences, topology, facts.
Smart ranking using relevance, recency, confidence, and importance.
Conflict detection and automatic reinforcement of related memories.
Built-in consolidation: merge duplicates, prune stale, promote frequent.
Local storage with SQLite and embeddings; no cloud.