Setting up MCP servers used to mean editing JSON config files, finding the right paths, and troubleshooting connection issues. Smithery changes everything.With one CLI command, you can auto-install and configure AI Memory MCP Server for Claude Desktop, Cursor, Windsurf, and 113+ other AI tools. Here's how.

What is Smithery?

Smithery is an open-source package manager and auto-configuration tool built specifically for the Model Context Protocol (MCP) ecosystem. Think of it as npm for MCP servers — but with a superpower: it doesn't just install packages, it automatically configures your AI clients to use them.

Before Smithery, setting up an MCP server required:

• Finding the correct configuration file for your AI client (e.g., claude_desktop_config.json for Claude Desktop)
• Writing the correct JSON configuration with the right command paths
• Restarting your AI client and hoping the connection works
• Repeating the process for each AI tool you use (Cursor, Windsurf, etc.)

With Smithery, you run one command, and it handles everything automatically.

Why Use Smithery for AI Memory MCP Server?

The AI Memory MCP Server gives your AI assistants persistent memory — the ability to search, save, and retrieve conversations across Claude, Cursor, Windsurf, and 113+ MCP clients. But setting it up manually can be tedious if you use multiple AI tools.

Smithery solves this by:

• Auto-detecting your installed MCP clients — Claude Desktop, Cursor, Windsurf, VS Code + Cline, and more
• Generating correct configuration — No JSON editing required
• Configuring all clients at once — One command sets up everything
• Verifying the connection — Confirms the server is working after install

Step-by-Step: Auto-Install AI Memory with Smithery

Step 1: Install the Smithery CLI (if needed)

The Smithery CLI runs via npx, so you don't need a separate install. Just run:

npx @smithery/cli --version

This confirms Smithery is available. If you don't have Node.js installed, download it here.

Step 2: Run the Auto-Install Command

Now run the Smithery install command for AI Memory:

npx @smithery/cli install aimemory-mcp-server

Smithery will:

1. Detect your MCP clients — Scans your system for Claude Desktop, Cursor, Windsurf, etc.
2. Ask which clients to configure — You can select all, or pick specific ones
3. Install the AI Memory MCP Server — Runs pip install aimemory-mcp-server automatically
4. Update each client's config file — Adds the correct JSON configuration
5. Verify the installation — Confirms each client can connect to the server

Step 3: Restart Your AI Clients

After Smithery finishes, restart each configured AI client:

• Claude Desktop: Quit and reopen the app
• Cursor: Reload the window (Cmd/Ctrl + Shift + P → "Reload Window")
• Windsurf: Restart the application

You should see "AI Memory" (or "aimemory-mcp-server") listed in the MCP servers section of your client.

Step 4: Test the Connection

Open a conversation in any configured client and try:

Search my memories for "database optimization"

Your AI should respond using the search_memoriestool from the AI Memory MCP Server. If it works, you're all set!

Manual Install vs. Smithery: Comparison

Configuring Specific AI Clients with Smithery

Claude Desktop + Smithery

Smithery auto-detects Claude Desktop's config file:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

After Smithery runs, the config file will include the AI Memory server entry. Restart Claude Desktop to see it in the MCP servers list.

Cursor + Smithery

Cursor's MCP settings are at Settings → MCP. Smithery automatically adds the AI Memory server to Cursor's MCP configuration. After restarting Cursor, you'll see the server in the MCP panel.

Windsurf + Smithery

Windsurf stores MCP config in its settings directory. Smithery detects and updates it automatically. Open Windsurf's settings to verify the AI Memory server is listed.

What MCP Tools Are Available After Install?

Once Smithery configures your AI clients, they gain access to 12 MCP tools from the AI Memory server:

• search_memories — Full-text search across all saved conversations
• save_memory — Save new memories, insights, or conversation snippets
• get_memory — Retrieve a specific memory by ID
• list_memories — Browse all memories with filtering and pagination
• update_memory — Edit existing memories to correct or expand them
• delete_memory — Remove outdated or irrelevant memories
• memory_stats — Get statistics about your memory store
• export_memories — Backup all memories to JSON
• import_memories — Import memories from JSON backup
• batch_save_memories — Save multiple memories at once
• get_all_tags — List all tags with usage counts
• clear_all_memories — Delete all memories (use with caution)

These tools work across all your MCP clients. Save a memory in Claude Desktop, retrieve it in Cursor — the memory store is shared.

Troubleshooting Smithery Install

"Command not found: npx"

You need Node.js installed. Download it from nodejs.org and retry.

"No MCP clients detected"

Make sure you've installed at least one MCP client (Claude Desktop, Cursor, or Windsurf). Smithery can only configure clients that are already installed.

"AI Memory tools not showing up"

Restart your AI client completely (quit and reopen). MCP servers are loaded on startup, so a restart is required.

"Connection refused" or "Server not responding"

Make sure the AI Memory MCP Server is installed: pip install aimemory-mcp-server. You can test it manually by running aimemory-mcp-server in your terminal.

Advanced: Using Smithery with Custom Configurations

If you need custom arguments or environment variables for the AI Memory server, you can edit the configuration after Smithery installs it. For example, to use a custom database path:

{
  "mcpServers": {
    "ai-memory": {
      "command": "aimemory-mcp-server",
      "args": [],
      "env": {
        "AIMEMORY_DB": "/path/to/custom.db"
      }
    }
  }
}

Smithery generates the base configuration, and you can customize it further if needed.

Conclusion: One Command, Persistent AI Memory

Smithery transforms MCP server setup from a manual, error-prone process into a one-command experience. With npx @smithery/cli install aimemory-mcp-server, you get:

• ✅ AI Memory MCP Server installed and running
• ✅ All your MCP clients auto-configured
• ✅ 12 memory tools available in Claude, Cursor, Windsurf, and 113+ other clients
• ✅ Persistent, searchable memory across all your AI conversations

Ready to give your AI persistent memory? Visit aimemory.pro to upload your existing conversations, or run the Smithery command above to get started in 30 seconds.

View on Smithery →