# MemOS

> Production-grade memory infrastructure for AI that provides scalable, persistent memory with millisecond-level response for intelligent applications.

MemOS is a memory-native framework designed to give AI systems continuous memory and growth capabilities. It provides production-grade memory infrastructure with millisecond-level response times, enabling AI applications to maintain consistent understanding and personalization across tasks and scenarios. The platform achieves SOTA performance on the LoCoMo Benchmark and supports enterprise partners including Trip, Anker, Haier, Alibaba Cloud, and China Telecom.

- **Structured Memory Architecture** unifies memory types into a layered system, enabling dynamic retrieval, updates, and smarter adaptive learning across AI applications
- **Millisecond Response Times** ensure each API call is stable and reliable, with fast and predictable responses for both add and search operations
- **Multi-scenario Deployment** supports public cloud, private cloud, on-premises, and hybrid architectures to meet needs from startups to large enterprises
- **Model-agnostic Compatibility** works with Agent frameworks, RAG setups, and major model ecosystems for seamless integration
- **Memory API Operations** provide full lifecycle memory control including CRUD, batch cleanup, tagging, and governance with consistency under heavy concurrency
- **Predictive Scheduling** employs intent-aware scheduling to preload relevant memory based on dialogue history, task semantics, or environmental cues
- **Memory Portability** enables memory sharing across models, devices, and apps via MIP protocol, making memory persistent and portable
- **Python SDK Integration** allows cloud integration in minutes with just a few lines of code for quick deployment

To get started, install the MemOS Python SDK using pip and initialize the client with your API key. The platform supports memory operations including adding messages, searching memory, and managing knowledge bases. MemOS is suitable for various intelligent application ecosystems including universal AI, companion apps, gaming, education, financial services, industrial applications, e-commerce, healthcare, legal, enterprise knowledge management, and customer support.

## Features
- Structured memory architecture
- Millisecond-level response times
- Memory API with CRUD operations
- Predictive and asynchronous scheduling
- Memory portability across models and devices
- Multi-scenario deployment support
- Model-agnostic compatibility
- Knowledge base management
- Chat API integration
- Python SDK
- Dynamic knowledge graph
- Batch cleanup and tagging
- Memory governance

## Integrations
Agent frameworks, RAG setups, Alibaba Cloud, Python applications

## Platforms
WINDOWS, WEB, API, DEVELOPER_SDK

## Pricing
Open Source, Free tier available

## Version
2.0

## Links
- Website: https://memos.openmem.net
- Documentation: https://memos-docs.openmem.net/
- Repository: https://github.com/MemTensor/MemOS
- EveryDev.ai: https://www.everydev.ai/tools/memos