MCP Memory

v1.0.0Knowledge & Memorystable

An advanced contextual memory server for agent orchestration providing chronological history, semantic vector search via Redis Stack, and temporary scratchpad storage. It also enables context compression through history rollups and facilitates inter-

cloudflarecloudflare-aicloudflare-d1cloudflare-vectorizecloudflare-worker
Share:
143
Stars
0
Downloads
0
Weekly
0/5

What is MCP Memory?

MCP Memory is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to advanced contextual memory server for agent orchestration providing chronological history, semantic vector search via redis stack, and temporary scratchpad storage. it also enables context compression...

An advanced contextual memory server for agent orchestration providing chronological history, semantic vector search via Redis Stack, and temporary scratchpad storage. It also enables context compression through history rollups and facilitates inter-

This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • An advanced contextual memory server for agent orchestration

Use Cases

An advanced contextual memory server for agent orchestration providing chronolog
Puliczek

Maintainer

LicenseMIT License
Languagehtml
Versionv1.0.0
UpdatedMay 12, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx mcp-memory

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use MCP Memory

MCP Memory is a persistent, vector-search-powered memory server for MCP clients such as Claude Desktop, Cursor, and Windsurf, allowing AI assistants to remember user preferences and past conversation context across sessions. It runs on Cloudflare Workers with Vectorize for semantic similarity search and D1 for relational storage, using the open-source bge-m3 embedding model via Workers AI. Memories are stored in isolated per-user namespaces, rate-limited to 100 requests per minute, and free to host for most personal usage levels on Cloudflare's free tier. A public hosted version is also available at memory.mcpgenerator.com for users who do not want to self-host.

Prerequisites

  • A Cloudflare account (free tier sufficient for personal use)
  • Node.js and npm installed for running Wrangler CLI
  • Wrangler CLI installed globally (npm install -g wrangler)
  • An MCP-compatible client such as Claude Desktop or Cursor
  • Optional: a userId string to namespace your memories (any unique identifier works)
1

Use the hosted version (quickest start)

Point your MCP client at the public hosted instance at memory.mcpgenerator.com. No deployment required. Skip to the client configuration step.

2

Deploy your own instance to Cloudflare (one-click)

Click the 'Deploy to Cloudflare' button in the GitHub repository, or use the Cloudflare CLI to create from the template. When prompted for Vectorize settings choose Dimensions: 1024 and Metric: cosine.

npm create cloudflare@latest --git https://github.com/puliczek/mcp-memory
3

Create the Vectorize index

After cloning or templating the repo, create the required vector index with the correct dimensions and distance metric.

npx wrangler vectorize create mcp-memory-vectorize --dimensions 1024 --metric cosine
4

Install dependencies and deploy

Install npm packages and deploy the Cloudflare Worker.

npm install
npm run deploy
5

Configure your MCP client

Add the deployed Worker URL (or the public hosted URL) to your MCP client configuration. Use your own unique userId to namespace your memories.

MCP Memory Examples

Client configuration

Claude Desktop configuration connecting to the public hosted MCP Memory instance. Replace the URL with your own deployed Worker URL if self-hosting.

{
  "mcpServers": {
    "mcp-memory": {
      "command": "npx",
      "args": ["mcp-memory"],
      "env": {
        "MCP_MEMORY_URL": "https://memory.mcpgenerator.com",
        "USER_ID": "your-unique-user-id"
      }
    }
  }
}

Prompts to try

Example prompts to store and retrieve memories once the server is connected.

- "Remember that I prefer TypeScript over JavaScript for all new projects"
- "What do you remember about my coding preferences?"
- "Store the fact that my timezone is UTC+5:30 and I work best in the morning"

Troubleshooting MCP Memory

Rate limit errors (429) when querying memories frequently

The default rate limit is 100 requests per minute per userId. If you hit it regularly, deploy your own instance and increase the rate limit in wrangler.jsonc. Self-hosted instances let you set any value you need.

Vectorize 'dimensions mismatch' error on deploy

The index must be created with exactly 1024 dimensions and cosine metric to match the bge-m3 model output. Delete any incorrectly configured index and recreate with: npx wrangler vectorize create mcp-memory-vectorize --dimensions 1024 --metric cosine.

Memories are not persisted between Cloudflare Worker requests

Ensure the Durable Object binding (MyMCP) is correctly configured in wrangler.jsonc and the D1 database binding is set. Run 'npx wrangler dev' locally to see binding errors before deploying.

Frequently Asked Questions about MCP Memory

What is MCP Memory?

MCP Memory is a Model Context Protocol (MCP) server that advanced contextual memory server for agent orchestration providing chronological history, semantic vector search via redis stack, and temporary scratchpad storage. it also enables context compression through history rollups and facilitates inter- It connects AI assistants to external tools and data sources through a standardized interface.

How do I install MCP Memory?

Follow the installation instructions on the MCP Memory GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with MCP Memory?

MCP Memory works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is MCP Memory free to use?

Yes, MCP Memory is open source and available under the MIT License license. You can use it freely in both personal and commercial projects.

Browse More Knowledge & Memory MCP Servers

Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "mcp-memory": { "command": "npx", "args": ["-y", "mcp-memory"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use MCP Memory?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides