Memory Graph Engine

v1.0.0Knowledge & Memorystable

Persistent memory graph for AI agents. Facts, decisions, entities, and relationships that survive across sessions, tools, and providers. MCP server — works with Claude, Cursor, ChatGPT, and any MCP client.

ai-agentsai-memoryai-toolsdeveloper-toolsembeddings
Share:
42
Stars
0
Downloads
0
Weekly
0/5

What is Memory Graph Engine?

Memory Graph Engine is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to persistent memory graph for ai agents. facts, decisions, entities, and relationships that survive across sessions, tools, and providers. mcp server — works with claude, cursor, chatgpt, and any mcp cl...

Persistent memory graph for AI agents. Facts, decisions, entities, and relationships that survive across sessions, tools, and providers. MCP server — works with Claude, Cursor, ChatGPT, and any MCP client.

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

Features

  • Persistent memory graph for AI agents. Facts, decisions, ent

Use Cases

Persistent agent memory
Fact and relationship storage
Cross-session continuity
kraklabs

Maintainer

LicenseAGPL-3.0
Languagego
Versionv1.0.0
UpdatedApr 24, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx mie

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 Memory Graph Engine

Mie (Memory Intelligence Engine) is a local-first MCP server written in Go that gives AI agents a persistent, typed knowledge graph — storing facts, decisions, entities, events, and relationships that survive across sessions, tools, and AI providers. It supports semantic search via embeddings (Ollama or OpenAI), multiple storage backends (RocksDB, SQLite, in-memory), and exposes twelve MCP tools for storing, querying, analyzing, and maintaining the graph. Developers and researchers use it to give any MCP-compatible AI client durable memory without relying on cloud storage.

Prerequisites

  • macOS with Homebrew, or Go 1.24+ if building from source
  • An MCP-compatible client such as Claude Code, Claude Desktop, or Cursor
  • Optional: Ollama running locally or an OpenAI API key if you want embedding-powered semantic search
  • Optional: RocksDB system libraries if using the rocksdb storage engine
1

Install Mie via Homebrew

Tap the kraklabs Homebrew repository and install the mie binary. Alternatively, build from source with Go 1.24+ if Homebrew is not available.

brew tap kraklabs/mie
brew install mie
2

Initialize the knowledge graph

Run mie init to create the default configuration and storage directory. Use the --interview flag for an interactive setup that configures storage engine and embedding provider.

# Default setup
mie init

# Interactive guided setup
mie init --interview
3

Configure storage and embeddings (optional)

Edit the generated YAML config to choose your preferred storage engine and embedding provider. SQLite is recommended for most users; RocksDB is better for large graphs.

storage:
  engine: sqlite
embedding:
  enabled: true
  provider: ollama
  model: nomic-embed-text
4

Register the server with your MCP client

Add a .mcp.json in your project root (for Claude Code) or edit claude_desktop_config.json (for Claude Desktop) with the mie --mcp command.

{
  "mcpServers": {
    "mie": {
      "command": "mie",
      "args": ["--mcp"]
    }
  }
}
5

Start the daemon and verify status

Start the mie daemon, then check the graph status to confirm the server is running and the storage backend is accessible.

mie daemon start
mie status

Memory Graph Engine Examples

Client configuration

Project-level .mcp.json entry for Claude Code using a locally installed mie binary.

{
  "mcpServers": {
    "mie": {
      "command": "mie",
      "args": ["--mcp"]
    }
  }
}

Prompts to try

Example prompts showing how to use Mie's persistent memory tools across sessions.

- "Remember that the production database host is db.prod.example.com and store it as a fact"
- "Store the decision we just made to use PostgreSQL over MySQL and why"
- "What do you remember about our authentication architecture?"
- "List all entities related to the payment service"
- "Check for any conflicting facts in the memory graph and report them"

Troubleshooting Memory Graph Engine

mie command not found after Homebrew install

Run 'brew link mie' and ensure /usr/local/bin or /opt/homebrew/bin is in your PATH. Restart your terminal and try 'which mie' to confirm the binary is visible.

Embeddings not working — semantic search returns empty results

Ensure your embedding provider is running. For Ollama, run 'ollama serve' and verify 'ollama list' shows the nomic-embed-text model. For OpenAI, confirm your API key is set in the config or environment.

RocksDB storage engine fails to start

RocksDB requires native system libraries. Switch to 'sqlite' in the storage.engine config field, or install rocksdb via 'brew install rocksdb' on macOS before retrying.

Frequently Asked Questions about Memory Graph Engine

What is Memory Graph Engine?

Memory Graph Engine is a Model Context Protocol (MCP) server that persistent memory graph for ai agents. facts, decisions, entities, and relationships that survive across sessions, tools, and providers. mcp server — works with claude, cursor, chatgpt, and any mcp client. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Memory Graph Engine?

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

Which AI clients work with Memory Graph Engine?

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

Is Memory Graph Engine free to use?

Yes, Memory Graph Engine is open source and available under the AGPL-3.0 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": { "mie": { "command": "npx", "args": ["-y", "mie"] } } }

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

Read the full setup guide →

Ready to use Memory Graph Engine?

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