Give your AI agent a permanent memory.

Developers add the SDK to their AI product. Their users get an assistant that actually remembers them. Free to start. No infrastructure needed.

Start Building FreeNo credit card required.

// 1) Initialize SDK client

import MemoryClient from "@youraimemory/sdk";

const client = new MemoryClient({ apiKey: process.env.YOURAIMEMORY_API_KEY });

// 2) Auto-extract from chat

await client.auto({

userId: "user_492",

messages: [

{ role: "user", content: "I prefer dark mode and Python." },

maxMemories: 5,

});

// 3) Recall relevant memories

const { memories } = await client.find({

userId: "user_492",

query: "theme preferences",

limit: 5,

});

// SDK response

{ "memories": [ { "content": "User prefers dark mode…" } ] }

Save Weeks

Skip building and maintaining complex vector databases.

3 Simple Calls

Store, search, and delete. A unified SDK designed for DX.

Works in 20 Min

Drop our SDK into your agent loop and ship today.

Free to Start

Generous free tier. Only pay when you scale to production.

SDK.SPECS.24

Core Operations

Everything you need to manage agent state.

SDK Method 01

Store Memory

Pass unstructured text. We handle chunking, embedding, and indexing automatically.

client.add({ userId, content })

-> { "id": "mem_123" }

SDK Method 02

Semantic Search

Retrieve perfectly formatted context arrays ready to inject into your LLM prompt.

client.find({ userId, query, limit: 5 })

-> [{ "content": "User loves coffee" }]

SDK Method 03

Manage & Delete

Delete by id (and list memories) to keep context relevant and compliant.

client.deleteMemory("mem_123")

-> 204 No Content

Build vs Buy

Stop reinventing the database infrastructure.

Building a scalable memory system requires managing vector DBs, embedding models, connection pooling, and latency optimization. We did that for you.

Save $12,000+ / yr

*Estimated engineering and infrastructure cost savings.

Component

Building In-House

With AI Memory

Infrastructure

Vector DB + Compute

Managed SDK + API

Setup Time

2–3 Weeks

20 Minutes

Maintenance

Ongoing scaling

Zero maintenance

Examples to get you started

See how to integrate AI Memory into tutor agents, chatbots, support agents, and more. Each example includes code explanation and full implementation.

Tutor Agent — How it works

After each session, extract memories from the conversation with client.auto(). Before the next lesson, recall them with client.find() and inject into your prompt.

// After session

await client.auto({ userId, messages });

// Before next lesson

const { memories } = await client.find({ userId, query: "weak topics" });

We have step-by-step examples for tutor agents, chatbots, support agents, and personal assistants. Each includes full code and integration guidance.

Browse all examples

Simple, transparent pricing.

Start for free. Scale when your agents get popular.

Hobby

$0/mo

Perfect for prototyping and personal projects.

SDK Calls10,000 / mo
Storage—
Community Support✓

Start Free