@memnexus-ai/sdk

Official Node.js / TypeScript SDK for the MemNexus API — persistent AI memory management.

Version: 1.53.28 Node.js: >= 18.0.0 Dependencies: none (uses native fetch)

Installation

npm install @memnexus-ai/sdk

Quick Start

import { MemNexus } from '@memnexus-ai/sdk';

const client = new MemNexus({ apiKey: 'your-api-key' });

// Save a memory
await client.remember('I deployed the new feature to production');

// Search memories
const results = await client.search('deployment history');
console.log(results.data);

// Build context briefing
const briefing = await client.brief('working on auth module');
console.log(briefing.data);

// Recall a digest
const digest = await client.recall('authentication decisions');
console.log(digest.data);

// Delete a memory
await client.forget('mem_abc123');

Authentication

Pass the API key directly or via the MEMNEXUS_API_KEY environment variable:

// Direct
const client = new MemNexus({ apiKey: 'your-api-key' });

// Environment variable
process.env.MEMNEXUS_API_KEY = 'your-api-key';
const client = new MemNexus();

Services

The client exposes a service for each API resource group:

const client = new MemNexus({ apiKey: 'your-api-key' });

// User management and profile endpoints
client.users

// Topic detection, clustering, and management endpoints
client.topics

// System health, monitoring, and configuration endpoints
client.system

// Pattern detection and behavioral analysis endpoints
client.patterns

// Behavioral pattern tracking and state management endpoints
client.behavior

// Behavioral recommendation management endpoints (v2)
client.recommendations

// Narrative thread management endpoints
client.narratives

// Observability and metrics endpoints for production monitoring
client.monitoring

// Memory management and retrieval endpoints
client.memories

// Invite code validation and gate status endpoints
client.invites

// Health check endpoints
client.health

// Graph-based retrieval augmented generation endpoints
client.graphrag

// Fact extraction and management endpoints
client.facts

// Entity extraction and discovery endpoints
client.entities

// Conversation tracking and analysis endpoints
client.conversations

// Subscription billing and payment management endpoints
client.billing

// Artifact storage and retrieval endpoints
client.artifacts

// API key management endpoints
client.apiKeys

// System alerts banner endpoints (customer portal)
client.alerts

// Admin management endpoints for invite codes and platform configuration
client.admin

// Admin endpoints for managing system alerts
client.adminAlerts

// Admin Pipeline
client.adminPipeline

Error Handling

import { MemNexus, SdkError } from '@memnexus-ai/sdk';

const client = new MemNexus({ apiKey: 'your-api-key' });

try {
  await client.remember('my memory');
} catch (err) {
  if (err instanceof SdkError) {
    console.error(`API error ${err.status}: ${err.statusText}`);
    console.error('Response body:', err.data);
  }
}

Configuration

const client = new MemNexus({
  apiKey: 'your-api-key',
  baseUrl: 'https://api.memnexus.ai',  // default
  timeoutMs: 30_000,         // default: 30 seconds
});

Retry Behaviour

The client automatically retries failed requests with exponential backoff:

• Max attempts: 5
• Base delay: 500 ms
• Max delay: 10 000 ms
• Backoff factor: 2×
• Retried status codes: 408, 429, 500, 502, 503, 504

License

Proprietary — Copyright © MemNexus Team