Memory infrastructure for AI agents

// npm install @scitrera/memorylayer-sdk
import { MemoryLayerClient, MemoryType } from '@scitrera/memorylayer-sdk';

// Connect to your local server
const ml = new MemoryLayerClient({
  baseUrl: 'http://localhost:61001',
  workspaceId: 'my-project'
});

// Store a memory
const memory = await ml.remember(
  'User prefers light mode with clean design',
  {
    type: MemoryType.SEMANTIC,
    importance: 0.8,
    tags: ['preferences', 'ui']
  }
);

// Recall memories
const memories = await ml.recall(
  'what are the user\'s UI preferences?'
);

// Synthesize insights
const reflection = await ml.reflect(
  'summarize all user preferences'
);

// Add to Claude Desktop config:
// ~/.config/claude/claude_desktop_config.json

{
  "mcpServers": {
    "memorylayer": {
      "command": "npx",
      "args": ["@scitrera/memorylayer-mcp-server"],
      "env": {
        "MEMORYLAYER_URL": "http://localhost:61001"
      }
    }
  }
}

// Available MCP tools include (see docs for more details):
// - memory_remember: Store new memories
// - memory_recall: Search memories
// - memory_reflect: Synthesize insights
// - memory_forget: Delete memories
// - memory_associate: Explicitly link memories with a relationship
// - memory_briefing: Session context
// - memory_session_start: Start a new session for working memory
// - memory_session_end: End session and optionally commit working memory
// - memory_session_status: Get current session status