Getting Started

Neuroloom gives your AI agents persistent, searchable memory. This guide walks through installing the SDK, authenticating, and creating your first memory.

Prerequisites

• Node.js 18.18+ or Python 3.11+
• A Neuroloom account — sign up free
• An API key (generated in the dashboard under Settings → API Keys)

Installation

Node.js / TypeScript

npm install @neuroloom/sdk

Python

pip install neuroloom

Quick Start

1. Initialize the client

import { Neuroloom } from "@neuroloom/sdk";

const client = new Neuroloom({
  apiKey: process.env.NEUROLOOM_API_KEY,
  workspaceId: process.env.NEUROLOOM_WORKSPACE_ID,
});

2. Create a memory

const memory = await client.memories.create({
  content: "The user prefers concise technical answers with code examples.",
  metadata: {
    category: "preferences",
    agentId: "assistant-v1",
  },
});

console.log(memory.id); // mem_01abc...

3. Search memories semantically

const results = await client.memories.search({
  query: "How does the user like to receive answers?",
  topK: 5,
});

results.forEach((r) => {
  console.log(r.content, r.score);
});

4. Retrieve a specific memory

const memory = await client.memories.get("mem_01abc...");
console.log(memory.content);

Environment Variables

Store your credentials in a .env file and never commit them to source control:

NEUROLOOM_API_KEY=nlk_live_...
NEUROLOOM_WORKSPACE_ID=ws_...

Next Steps

• Authentication — Learn about API key scopes and workspace isolation
• Memories API — Full reference for create, search, update, and delete operations
• MCP Integration — Connect Neuroloom to Claude Desktop and other MCP clients