@youraimemory/sdk

A tiny TypeScript/JavaScript client for the YourAIMemory API. It makes it easy to store, search, and auto-extract memories about your users with just a few lines of code.

Installation

npm install @youraimemory/sdk
# or
yarn add @youraimemory/sdk
# or
pnpm add @youraimemory/sdk

Quick start

import MemoryClient from "@youraimemory/sdk";

const client = new MemoryClient({
  // If omitted, the client will fall back to process.env.YOURAIMEMORY_API_KEY (in Node)
  apiKey: process.env.YOURAIMEMORY_API_KEY,
});

// Add a memory
await client.add({
  userId: "user_123",
  content: "Meeting notes from Q1 planning",
});

// Find memories
const result = await client.find({
  userId: "user_123",
  query: "planning notes",
  // optional:
  // limit: 10,
  // relevanceThreshold: 0.7,
});

console.log(result); // { memories: [...] }

API Overview

The client exposes a small, ergonomic surface on top of the HTTP API:

• new MemoryClient(options)

  • apiKey?: string – Your YourAIMemory API key. If omitted (Node), falls back to process.env.YOURAIMEMORY_API_KEY.
  • baseUrl?: string – Override the API base URL (defaults to https://api.youraimemory.com).
  • fetchFn?: typeof fetch – Custom fetch implementation (for tests, Node polyfills, etc.).

• High-level helpers

  • client.add({ userId, content, metadata? }) Store a single memory for a user.
  • client.find({ userId, query, limit?, relevanceThreshold? }) Search for the most relevant memories for a user.
  • client.auto({ userId, messages, maxMemories?, importanceThreshold? }) Send a conversation transcript and let the API extract and store memories automatically.
  • client.list({ userId, limit?, offset? }) List recent memories for a user.
  • client.deleteMemory(memoryId) Delete a memory by its ID.

• Low-level access

  • client.memories.auto(...)
  • client.memories.search(...)
  • client.memories.store(...)
  • client.memories.list(...)
  • client.memories.delete(...)

These map 1:1 to the underlying /v1/memories HTTP endpoints and accept the full typed request payloads from @aimemory/shared.

Why use YourAIMemory?

• Long‑term memory for your AI apps Store user profile details, preferences, tasks, and episodic events so your AI can stay consistent over time.

• Smart retrieval Hybrid search (vectors + keywords), recency, importance, usage frequency, and memory type signals are combined to surface the most relevant memories.

• Automatic extraction Send raw conversation logs and let the API extract structured memories for you.

• Simple integration Just a few lines of code with client.add and client.find – no need to manage embeddings, indexes, or ranking logic yourself.