@mnemoai/client

Official client for Mnemo Cloud — long-term memory for AI agents.

Install

npm install @mnemoai/client

Quick Start

import { createCloudMnemo } from "@mnemoai/client";

const mnemo = createCloudMnemo({ apiKey: "mn_your_key" });

// Store a memory
const { id } = await mnemo.store({
  text: "User is a software engineer who prefers TypeScript",
  category: "fact",
  importance: 0.9,
});

// Recall memories by semantic search
const memories = await mnemo.recall("what does the user do?", { limit: 5 });
memories.forEach(m => console.log(m.text, m.score));

// Delete a memory
await mnemo.delete(id);

// Check usage
const stats = await mnemo.stats();
console.log(stats.totalEntries, "memories stored");

API

createCloudMnemo(config)

| Parameter | Type | Required | Description | |-----------|------|----------|-------------| | apiKey | string | Yes | API key from Mnemo Cloud | | endpoint | string | No | Custom endpoint (default: https://api.m-nemo.ai) |

mnemo.store(entry)

Store a memory. Returns { id: string }.

| Field | Type | Default | Description | |-------|------|---------|-------------| | text | string | required | The text to remember | | category | string | "fact" | fact, preference, decision, entity, other | | importance | number | 0.7 | Importance score (0.0–1.0) | | scope | string | "global" | Scope for multi-agent isolation |

mnemo.recall(query, options?)

Recall memories by semantic search. Returns array of { text, score, category, importance, timestamp }.

| Option | Type | Default | Description | |--------|------|---------|-------------| | limit | number | 5 | Max results | | category | string | — | Filter by category | | scopeFilter | string[] | — | Filter by scopes |

mnemo.delete(id)

Delete a memory by ID. Returns boolean.

mnemo.stats()

Returns { totalEntries, scopeCounts, categoryCounts, plan }.

Requirements

• Node.js 18+ (uses native fetch)
• Mnemo Cloud API key

Links

• Mnemo Cloud docs
• Full API reference
• GitHub