# Product updates, implementation guides, and notes from the Mnexium team.

Learn how to build memory-aware AI products with durable context, structured data, and production-grade agent infrastructure.

## [Introducing Integrations: Bring Live External Data Into Mnexium](/content/blogs/introducing-integrations/index.html)

A deep dive into Mnexium Integrations: why they matter, how pull and webhook modes work, and how teams can inject live operational data into runtime prompts.

## [Memory Policies: Control What Your AI Remembers](/content/blogs/memory-policies-docs/index.html)

A deeper look at Memory Policies in Mnexium: why they matter, how scoped defaults work, and what teams can accomplish by controlling extraction quality.

## [Persistent Memory API for AI Apps (OpenAI, Claude, Gemini)](/content/blogs/persistent-memory-api/index.html)

Provider hub page for persistent memory APIs, with direct guides for OpenAI, Claude, and Gemini implementations.

## [Persistent Memory API for Gemini Apps](/content/blogs/persistent-memory-api-gemini/index.html)

Intent guide for Google Gemini teams: add durable memory, context continuity, and history with Mnexium.

## [Persistent Memory API for Claude Apps](/content/blogs/persistent-memory-api-claude/index.html)

Intent guide for Anthropic Claude teams: add long-term memory and session continuity with Mnexium.

## [Persistent Memory API for OpenAI Apps](/content/blogs/persistent-memory-api-openai/index.html)

Intent guide for teams building on OpenAI: add durable memory, chat continuity, and structured context without building a custom memory stack.

## [Introducing Cartly: An iOS Receipt Tracking App Built on Mnexium](/content/blogs/introducing-cartly/index.html)

Feature-focused breakdown of the Mnexium runtime in Cartly: mnx controls, durable identity, memory flags, records schemas, and deterministic OCR sync.

## [Introducing the Mnexium n8n Connector](/content/blogs/introducing-mnexium-n8n-connector/index.html)

Production-ready Mnexium integration for n8n with chat, memory, profile, records, and custom API operations.

## [Introducing Memory Policies: Controlled Memory Extraction for AI](/content/blogs/introducing-memory-policies/index.html)

Memory Policies let you control extraction behavior with scoped defaults and per-request overrides, so your assistant stores high-signal memories with predictable rules.

## [Introducing Records: Structured Data for AI Applications](/content/blogs/introducing-records/index.html)

Records give your AI a structured data layer — define schemas, CRUD records via API, query with filters, search semantically, and let the AI create and update records automatically from conversations.

## [Introducing the Mnexium SDKs for JavaScript and Python](/content/blogs/introducing-sdks/index.html)

Official SDKs for JavaScript/TypeScript and Python. Add persistent memory, user profiles, structured claims, and conversation history to your AI applications in five lines of code.

## [Introducing @mnexium/chat: AI Chat for Any Website](/content/blogs/introducing-mnexium-chat/index.html)

A single npm package that adds a polished, production-ready AI chat widget to any website. React, Next.js, Express, or plain HTML — it just works, and most importantly, it remembers.

## [Memory Decay: How AI Remembers Like Humans Do](/content/blogs/memory-decay/index.html)

Introducing Memory Decay — a system that makes AI memory behave more like human memory. Frequently used memories stay strong, while unused ones naturally fade. The result? More relevant, contextual AI interactions.

## [Chat Summarization: Cut Your Token Costs by 95%](/content/blogs/chat-summarization/index.html)

Long conversations are expensive. Chat Summarization intelligently compresses your conversation history while preserving context, slashing your token costs dramatically with rolling summarization.

## [Introducing Profiles: Structured User Data for AI](/content/blogs/introducing-profiles/index.html)

Profiles give your AI applications instant access to key user attributes like names, emails, timezones, and custom fields — all automatically extracted from conversations or set via API.

## [Hello Mnexium! Getting Started](/content/blogs/hello-mnexium/index.html)

In this getting started Hello World post - we clone chatGPT using Mnexium. We include memory and conversation history.