@yarlisai/memory

AI memory layer with pluggable backends for the Yarlis AI ecosystem. Store, query, and inject contextual memories into AI conversations.

Install

npm install @yarlisai/memory @yarlisai/contracts

Quick Start

import { YarlisMemory, InMemoryStore } from "@yarlisai/memory";
import { createYPID } from "@yarlisai/contracts";

const memory = new YarlisMemory(new InMemoryStore());
const userId = createYPID("rtm", "user");

// Store a memory
await memory.remember(userId, "User prefers dark mode", {
  scope: "user",
  type: "semantic",
  domain: "rtm",
});

// Recall memories
const memories = await memory.recall(userId, { scope: "user" });

// Inject memories into a conversation
const messages = [{ role: "user" as const, content: "What are my preferences?" }];
const enriched = await memory.injectContext(userId, messages);

Adapters

• InMemoryStore — In-memory store for development and testing
• FirestoreMemoryStore — Google Firestore backend for production Firebase projects
• SQLiteMemoryStore — Local persistent SQLite backend

License

MIT