@RookieRicardoR: MemOS has made new progress. There are quite a few AI Memory solutions now, but many are still at the level of storing chat history. It looks like memory, but it's essentially adding a semantic search to markdown. @MemOS_dev has been working on a memory system for a while, from 1.0 all the way...

MemOS has made another leap forward.

There are quite a few approaches to AI Memory now, but many just stop at storing chat logs on the surface — it looks like memory, but in reality it’s just adding semantic search to Markdown.

@MemOS_dev has been working on memory systems for a while, evolving from 1.0 all the way to 2.0. Looking back, a consensus is slowly becoming clear: the next thing that really determines whether an Agent is useful isn’t the model parameters themselves — it’s whether the Agent can continuously learn in your own local world, accumulate experience, remember feedback, and reuse capabilities.

Recently, after the official launch of MemOS Local Plugin 2.0, I took a deep dive into its new features.

The feature I find most worth talking about in this update is called “Execute to Learn.”

It doesn’t just store conversations. Instead, when the Agent helps you complete a task, it breaks down the key steps in the execution process into learnable units, then transforms them into cognitive assets that can be scored, attributed, and reused.

In simpler terms: it doesn’t just remember what you said — it remembers how you and the Agent completed a task step by step.

Let me share a recent personal experience. When I was writing a small tool with OpenClaw, in the first round I laid out the code style, naming conventions, and error handling approach one by one. It took three or four rounds of back-and-forth to get it right.

In the past, switching to a new conversation or coming back a couple of days later would basically lose all that context — I’d have to explain everything from scratch again. But with this 2.0 version, when I started a new task the next day, it automatically applied the same patterns we had refined in the previous session.

Let me walk through the new features one by one:

https://github.com/MemTensor/MemOS/tree/main/apps/memos-local-plugin…