Documentation

Step-by-step guide to adding persistent memory to your AI assistant.

Quick Start

Get memory working in under a minute. No signup required for local mode.

bash

npx @lakehouse42/memory-mcp

This starts the MCP server with local storage. Your memories are saved to ~/.lakehouse/memory-mcp/memories.json

Claude Code

Add memory to Claude Code by editing your config file.

Step 1: Edit config

Open ~/.claude/claude_code_config.json

~/.claude/claude_code_config.json

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": ["@lakehouse42/memory-mcp"]
    }
  }
}

Step 2: Restart Claude Code

Close and reopen your terminal. The memory tools will now be available.

Tip: Ask Claude to remember that I prefer TypeScript to test that memory is working.

Claude Desktop

Add memory to Claude Desktop via the settings menu.

1Open Claude Desktop Settings
2Go to Developer → Edit Config
3Add the MCP server configuration

Claude Desktop Config

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": ["@lakehouse42/memory-mcp"],
      "env": {
        "LAKEHOUSE_URL": "https://api.lakehouse42.com",
        "LAKEHOUSE_API_KEY": "lh_your_api_key"
      }
    }
  }
}

Cursor / Windsurf

Add memory to Cursor or Windsurf following the same pattern.

MCP Config

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": ["@lakehouse42/memory-mcp"]
    }
  }
}

LH42 Backend

Connect to LH42 for full semantic search, deduplication, and knowledge graph features.

Get your API key

Sign up for LH42
Go to Settings → API Keys
Create a new API key

Add environment variables

config.json

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": ["@lakehouse42/memory-mcp"],
      "env": {
        "LAKEHOUSE_URL": "https://api.lakehouse42.com",
        "LAKEHOUSE_API_KEY": "lh_your_api_key"
      }
    }
  }
}

Variable	Description
LAKEHOUSE_URL	LH42 API URL (enables cloud backend)
LAKEHOUSE_API_KEY	API key for authentication
DEBUG	Enable debug logging (true/false)

MCP Tools Reference

remember

Store a memory for later recall.

Parameter	Type	Description
content*	string	The information to remember
type	string	Memory type (fact, preference, task, event, context, reflection)
importance	number	0.0 to 1.0 (default: 0.5)

recall

Search memories by semantic similarity.

Parameter	Type	Description
query*	string	What to search for
limit	number	Max results (default: 5)
types	string[]	Filter by memory types

forget

Delete a memory by ID.

Parameter	Type	Description
memoryId*	string	ID of memory to delete
reason	string	Reason for deletion (audit trail)

list_memories

List recent memories.

Parameter	Type	Description
limit	number	Max results (default: 10)

memory_status

Get memory system status including backend type and available features.

Memory Types

Type	Description	Example
fact	Factual information (default)	"The API uses REST endpoints"
preference	User preferences	"User prefers dark mode"
task	Tasks and todos	"Need to refactor the auth module"
event	Events and occurrences	"Deployed v2.0 on March 15"
context	Conversation context	"Working on the payment flow"
reflection	Insights and learnings	"The old approach was too slow"

REST API

Access memory directly via the LH42 REST API.

bash

# Store a memory
curl -X POST https://api.lakehouse42.com/api/v1/memory \
  -H "X-API-Key: lh_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{"content": "User prefers TypeScript", "memory_type": "preference"}'

bash

# Search memories
curl "https://api.lakehouse42.com/api/v1/memory/search?query=preferences" \
  -H "X-API-Key: lh_your_api_key"

Full API documentation

Programmatic Usage

Use the package in your own TypeScript/JavaScript code.

typescript

import { createMemoryServer, LakehouseBackend } from "@lakehouse42/memory-mcp";

// Create server with custom config
const server = await createMemoryServer({
  lakehouseUrl: "https://api.lakehouse42.com",
  apiKey: "lh_xxx",
  debug: true,
});

// Or use backends directly
const backend = new LakehouseBackend({
  url: "https://api.lakehouse42.com",
  apiKey: "lh_xxx",
});

await backend.initialize();
await backend.remember({ content: "User likes TypeScript" });
const results = await backend.recall({ query: "programming preferences" });

Ready to upgrade?

Get full semantic search, deduplication, and knowledge graph features with LH42.