Nocturne Memory

v1.0.0Knowledge & Memorystable

A lightweight, rollbackable, and visual Long-Term Memory Server for MCP Agents. Say goodbye to Vector RAG and amnesia. Empower your AI with persistent, graph-like structured memory across any model, session, or tool. Drop-in replacement for OpenClaw.

agentic-aiai-identityai-memoryartificial-intelligenceclaude
Share:
1,120
Stars
0
Downloads
0
Weekly
0/5

What is Nocturne Memory?

Nocturne Memory is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to lightweight, rollbackable, and visual long-term memory server for mcp agents. say goodbye to vector rag and amnesia. empower your ai with persistent, graph-like structured memory across any model, ses...

A lightweight, rollbackable, and visual Long-Term Memory Server for MCP Agents. Say goodbye to Vector RAG and amnesia. Empower your AI with persistent, graph-like structured memory across any model, session, or tool. Drop-in replacement for OpenClaw.

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

Features

  • A lightweight, rollbackable, and visual Long-Term Memory Ser

Use Cases

Persistent graph-based memory for AI agents
Long-term recall across sessions and contexts
Dataojitori

Maintainer

LicenseMIT
Languagepython
Versionv1.0.0
UpdatedMay 21, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx nocturne-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 Nocturne Memory

Nocturne Memory is a lightweight, graph-based long-term memory server for MCP agents that stores, links, and retrieves AI memories across sessions, model switches, and tool boundaries. Unlike vector RAG systems, it uses structured hyperlinking with Aho-Corasick keyword matching to connect related memories automatically, and provides a web dashboard for visual browsing and point-in-time rollback. It is designed as a drop-in replacement for OpenClaw memory and supports namespace isolation so multiple distinct AI personas can share a single database without interference.

Prerequisites

  • Python 3.10 or later installed
  • Node.js installed (the server auto-builds the frontend on first launch)
  • An MCP-compatible client such as Claude Desktop, Cursor, or Gemini CLI
  • Optional: PostgreSQL for production deployments (SQLite is used by default)
  • Git for cloning the repository
1

Clone the repository

Clone the Nocturne Memory repository from GitHub to your local machine.

git clone https://github.com/Dataojitori/nocturne_memory.git
cd nocturne_memory
2

Install Python dependencies

Install the required Python packages listed in the backend requirements file.

pip install -r backend/requirements.txt
3

Configure your MCP client

Add Nocturne Memory as an MCP server in your AI client's configuration. Use the absolute path to the mcp_server.py file for your system.

{
  "mcpServers": {
    "nocturne_memory": {
      "command": "python",
      "args": ["/absolute/path/to/nocturne_memory/backend/mcp_server.py"]
    }
  }
}
4

Verify the connection

After connecting your MCP client, instruct the AI to read the boot memory to confirm the server is running and the agent can access its memories.

5

Open the visual dashboard

The server auto-builds and serves a web UI on first launch. Open it in your browser to visually browse the memory graph, inspect individual memories, and roll back to previous snapshots if needed.

# The dashboard is served at http://localhost:PORT after the server starts
# Check server output for the exact port number
6

Configure namespaces for multiple personas

Set the NAMESPACE environment variable in your MCP server config to isolate memories for different AI personas or projects. Each namespace operates as an independent memory space within the same database.

{
  "mcpServers": {
    "nocturne_memory": {
      "command": "python",
      "args": ["/path/to/nocturne_memory/backend/mcp_server.py"],
      "env": {
        "NAMESPACE": "research-agent"
      }
    }
  }
}

Nocturne Memory Examples

Client configuration

Claude Desktop configuration for Nocturne Memory with a custom namespace for isolation.

{
  "mcpServers": {
    "nocturne_memory": {
      "command": "python",
      "args": ["/Users/you/nocturne_memory/backend/mcp_server.py"],
      "env": {
        "NAMESPACE": "default",
        "DATABASE_URL": "sqlite:///memories.db"
      }
    }
  }
}

Prompts to try

Example prompts that exercise Nocturne Memory's read, create, search, and rollback capabilities.

- "Read system://boot and tell me who you are and what you remember"
- "Create a memory in the 'research' domain called 'project-goals' with my current objectives"
- "Search memory for everything related to 'machine learning'"
- "What are the most recent memories you have? Read system://recent"
- "Add an alias 'ml-notes' to the memory about machine learning so I can access it faster"

Troubleshooting Nocturne Memory

The MCP client cannot connect or the server exits immediately on startup

Verify the absolute path to mcp_server.py is correct and that Python 3.10+ is the interpreter being invoked. Run 'python /path/to/backend/mcp_server.py' directly in your terminal to see startup errors. Also confirm all requirements from backend/requirements.txt are installed in the same Python environment.

The frontend dashboard fails to load or the auto-build step fails

Ensure Node.js is installed ('node --version') since the server auto-compiles the frontend on first launch. If the build fails, set the environment variable SKIP_FRONTEND_BUILD=true in the server config and access the API directly. You can also manually build the frontend by running 'npm install && npm run build' in the frontend directory.

Memories from a previous session are not accessible

Confirm the NAMESPACE value in your MCP config matches the one used during the previous session. Also verify the DATABASE_URL points to the same database file or server used previously. Memories are namespace-scoped, so switching namespaces will show a different memory space.

Frequently Asked Questions about Nocturne Memory

What is Nocturne Memory?

Nocturne Memory is a Model Context Protocol (MCP) server that lightweight, rollbackable, and visual long-term memory server for mcp agents. say goodbye to vector rag and amnesia. empower your ai with persistent, graph-like structured memory across any model, session, or tool. drop-in replacement for openclaw. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Nocturne Memory?

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

Which AI clients work with Nocturne Memory?

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

Is Nocturne Memory free to use?

Yes, Nocturne Memory is open source and available under the MIT 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": { "nocturne-memory": { "command": "npx", "args": ["-y", "nocturne-memory"] } } }

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

Read the full setup guide →

Ready to use Nocturne 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