Add Persistent Memory to Claude Code in 5 Minutes with memoclaw-mcp

What It Does — Persistent Memory as an MCP Tool

Your Claude Code agent forgets everything between sessions. You've likely hacked around this with MEMORY.md files or by re-stating your preferences in every new chat. The memoclaw-mcp server fixes this by adding semantic memory-as-a-service directly to any MCP-compatible client, including Claude Code. It exposes three native tools—store_memory, recall_memories, and list_memories—that let your agent remember project details, your coding preferences, and architectural decisions across sessions.

Setup — How to Install and Configure with Claude Code

Installation is a single command. You'll need Node.js 18+ and an Ethereum wallet (like MetaMask) for identity; no API keys or account registration is required.

npm install -g memoclaw-mcp

Next, configure Claude Code to use the server. Find or create your Claude Desktop configuration file (typically claude_desktop_config.json) and add the MemoClaw server block.

{
  "mcpServers": {
    "memoclaw": {
      "command": "memoclaw-mcp",
      "env": {
        "MEMOCLAW_PRIVATE_KEY": "your-wallet-private-key"
      }
    }
  }
}

Restart Claude Code. That's it—your agent now has access to persistent memory tools.

When To Use It — Specific Use Cases Where It Shines

This server eliminates the need to manually manage context. Use it to:

• Remember Personal Preferences: Tell Claude Code once that you prefer TypeScript, and it will recall this for all future project setups.

  You: "Remember that I prefer TypeScript over JavaScript for all new projects and use pnpm as my package manager."
  Agent calls store_memory with tags ["preference", "language", "tooling"].

• Maintain Project Context: Store key architectural decisions or client requirements that persist beyond a single coding session.

  // Example of storing a project-specific memory
  store_memory({
    content: "The /api/v2 endpoint uses Bearer token auth, not API keys.",
    importance: 0.9,
    namespace: "project-saas-backend",
    tags: ["api", "authentication", "architecture"]
  })
  

• Isolate Memories with Namespaces: Work on multiple projects without cross-contamination. Use the namespace parameter (like "project-acme") to keep memories for different codebases completely separate.

When you start a new session and ask, "What are the auth rules for the SaaS backend?", your agent can call recall_memories({ query: "authentication api", namespace: "project-saas-backend" }) and get the exact note you stored last week.