MemMachine

MemMachine is an open-source long-term memory layer for AI agents and LLM-powered applications, released under the Apache 2.0 license. Initially incubated by MemVerge, the project is now community-driven and provides developers with a model-agnostic memory infrastructure that persists across sessions, agents, and LLM providers. It is available as a self-hosted deployment or via a managed cloud platform.

What It Is

MemMachine gives AI agents the ability to learn, store, and recall information from past interactions—transforming stateless chatbots into personalized, context-aware assistants. Unlike memory services tied to a single frontier model provider, MemMachine is designed to work with any LLM, including OpenAI, Anthropic, Bedrock, Ollama, and privately hosted models. This model-agnostic design lets organizations avoid vendor lock-in and maintain full control over their data.

Memory Architecture

MemMachine implements three distinct memory types, each backed by a purpose-fit storage layer:

• Working Memory: Short-term context for the current session
• Episodic Memory: Graph-based (Neo4j) conversational context that persists across sessions, capturing the history and flow of interactions
• Profile Memory: Long-term user facts and preferences stored in a SQL database, enabling true personalization over time

Agents interact with the memory core through a RESTful API, a Python SDK, a TypeScript SDK, or a native MCP (Model Context Protocol) server. The MCP server supports both stdio mode (for Claude Desktop) and HTTP mode (for web clients).

Integrations and Ecosystem

MemMachine ships with first-class integrations for major AI frameworks:

• LangChain – memory provider for LangChain agents
• LangGraph – stateful memory for LangGraph workflows
• CrewAI – persistent memory for multi-agent systems
• LlamaIndex – memory integration for LlamaIndex applications
• AWS Strands Agent SDK – memory for AWS Strands agents
• n8n – no-code workflow automation integration
• Dify and FastGPT – memory backends for those platforms

The project also includes a blog post describing enterprise-grade memory capabilities available for the NVIDIA NeMo Agent Toolkit, published in April 2026.

Deployment Model

MemMachine can be run locally via Docker, self-hosted in a private cloud or on-premises data center, or accessed through the managed MemMachine Platform at console.memmachine.ai. The self-hosted path gives organizations full data sovereignty, which the FAQ highlights as a key differentiator from frontier-lab memory services that retain user data on their own infrastructure.

Getting started requires installing the memmachine-client Python package and pointing it at a running MemMachine server. The README describes the setup as achievable in under five minutes.

Update: v0.3.9

The latest GitHub release is v0.3.9, published May 18, 2026. The repository was created in August 2025 and has seen active development, with the last push recorded in September 2026. The project has accumulated over 3,200 GitHub stars and 215 forks since launch, signaling growing community interest. An Enterprise version with additional features and dedicated support is noted on the pricing page as coming soon.