Claude Code Memory Server

You are deploying a Neo4j-backed MCP server that gives Claude Code persistent memory across tasks and projects. It tracks tasks, patterns, decisions, and relationships to enable contextual development assistance, intelligent search, and workflow optimization across sessions.

How to use

You interact with the Claude Code Memory MCP server through an MCP client to store, retrieve, and analyze memories. You can persist development tasks, code patterns, problems, solutions, and project context, then query the memory graph to discover related concepts, reuse patterns, and optimal workflows. Use it to track task execution, compare past approaches, and surface contextual knowledge when working in a new project or technology.

How to install

Prerequisites: ensure you have Python 3.10 or newer and a running Neo4j instance (local or cloud). You will also need access to Claude Code with MCP support to connect to the memory server.

Step 1: Set up your environment and clone the project files.

# Clone the memory MCP repository
git clone https://github.com/viralvoodoo/claude-code-memory.git
cd claude-code-memory

Step 2: Install dependencies for the memory MCP server.

pip install -e .

Step 3: Configure your Neo4j connection details.

cp .env.example .env

Edit the environment file to include your Neo4j credentials, such as the database URI, username, and password.

Step 4: Initialize the database schema used by the memory graph.

python -m claude_memory.setup

Step 5: Start the MCP server via the provided runtime command.

python -m claude_memory.server

## Configuration and connection tokens

The server uses a Neo4j graph database to store memories and relationships. The following environment variables configure the connection and logging:

## Environment variables

- NEO4J\_URI: Neo4j database URI (default: bolt://localhost:7687)
- NEO4J\_USER: Database username (default: neo4j)
- NEO4J\_PASSWORD: Database password
- MEMORY\_LOG\_LEVEL: Logging level (default: INFO)

## MCP client integration example

To integrate Claude Code with the memory MCP server, configure the MCP client to point to the local Python runtime that runs the server.

## Usage with the Claude Code memory MCP client

Once connected, you can perform core memory operations, manage relationships, and query the knowledge graph to extract patterns and guidance for ongoing development tasks.

## Additional notes

Refer to the setup steps for a clean start, and ensure your Neo4j instance is reachable from the machine running the MCP server. Use secure credentials and rotate them periodically to maintain security.

## Security considerations

Run the memory server within a trusted network, restrict access to the Neo4j instance, and monitor logs for unusual activity. Use strong passwords and consider enabling encryption for database connections.

## Troubleshooting

If the server fails to start, verify that Python can locate the claude\_memory module, confirm Neo4j connectivity, and check that the .env file contains correct credentials. Review the log level for more detailed output.

## Available tools

### store\_memory

Store new development memories with context to build a persistent knowledge base.

### get\_memory

Retrieve a specific memory by ID along with its relationships.

### search\_memories

Find memories by content, context, or relationships across the graph.

### update\_memory

Modify existing memory content and metadata.

### delete\_memory

Remove a memory and cleanup its relationships.

### create\_relationship

Link memories with specific relationship types to build a rich graph.

### get\_related\_memories

Find memories connected to a particular memory through relationships.

### analyze\_relationships

Identify patterns in relationships to reveal insights.

### analyze\_codebase

Scan a project and create a contextual memory graph.

### track\_task\_execution

Record development workflows and observed patterns.

### suggest\_similar\_solutions

Identify past solutions similar to the current challenge.

### predict\_solution\_effectiveness

Estimate the likelihood of success for proposed approaches.

### get\_memory\_graph

Visualize the knowledge network and all relationships.

### find\_memory\_paths

Discover connection chains between concepts in the graph.

### memory\_effectiveness

Track and analyze the success rates of solutions over time.