MCP Memory Visualizer

MCP Memory Visualizer is a visualization utility that maps and analyzes memory data generated by Anthropic's Memory MCP server. Designed for developers, researchers, and users of Claude Desktop, it converts stored knowledge graph records into visual representations. The server works directly with the memory.json file produced by the official memory server, interpreting entity nodes, observation lists, and relational edges. Users can view their data using an in-browser web app, a static NetworkX script generating high-resolution graphs, or an interactive PyVis script creating local HTML files. By surfacing connection patterns, centrality measurements, and isolated nodes, MCP Memory Visualizer allows users to understand how Claude organizes context across sessions. It helps users audit what an assistant retains, diagnose sparse or redundant entities, and optimize long-term memory configurations without sending sensitive local data to third-party endpoints.

Category: AI Memory & Context

Tags: analysis, claude, graph, memory, visualization

Visit MCP Memory Visualizer

How to install and configure MCP Memory Visualizer

1. Configure your Claude Desktop configuration file to store memory in a persistent path using the MEMORY_FILE_PATH environment variable: json { "mcpServers": { "memory": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-memory"], "env": { "MEMORY_FILE_PATH": "C:\\Users\\<username>\\Documents\\claude-memory\\memory.json" } } } } 2. To use the browser version, open the interactive web visualizer at https://dzivkovi.github.io/mcp-memory-visualizer/ and drag and drop your memory.json file. 3. Alternatively, clone the repository and install the Python dependencies: bash pip install -r requirements.txt 4. Run python visualize_memory.py for static graph generation or python visualize_memory_interactive.py for interactive PyVis browser output.

What you can do with MCP Memory Visualizer

• Mapping relationships and entity networks created by Claude Desktop across multiple past chat conversations. - Identifying redundant, duplicate, or sparse memory nodes to clean up and optimize the memory.json file. - Generating high-resolution network diagrams and statistical summaries for AI memory and knowledge retrieval research. - Inspecting stored observations and connection densities locally without uploading sensitive files to external servers.

Key facts

• https://github.com/dzivkovi/mcp-memory-visualizer
• AI Memory & Context, Design, Media & Creative
• analysis, claude, graph, memory, visualization

Part of MCP Servers

What can MCP Memory Visualizer do?

MCP Memory Visualizer renders graph diagrams from Claude memory.json files created by the official Memory MCP server. It provides force-directed interactive browser graphs, static NetworkX network plots, centrality metrics, and redundancy detection to help users inspect and optimize their AI's persistent context.

How do I install MCP Memory Visualizer?

You can use the tool without installation by uploading your memory.json file directly to the web visualizer. For local Python execution, clone the repository, install dependencies with pip install -r requirements.txt, and run either visualize_memory.py or visualize_memory_interactive.py.

Where does Claude store the memory.json file?

By default, the Memory MCP server places memory.json inside a temporary npm cache folder that can be wiped during cache cleans. Anthropic configuration recommends setting the MEMORY_FILE_PATH environment variable in your Claude Desktop configuration to a stable location, such as your Documents directory.

Is data uploaded to external servers when visualizing memory?

No. The web-based visualizer processes files entirely client-side within your browser using D3.js. The Python scripts also execute entirely on your local machine using NetworkX and PyVis, ensuring memory contents remain private.