Continuity layer vs agent memory?

[Alake-SwordFish]

Hi, I found ReMe’s work on stateless sessions and portable memory really interesting.
I’m exploring a concept called the Presence Continuity Layer, basically continuity for the user across models, devices, and sessions, rather than memory only inside the agent.
Does that feel like a meaningful distinction to you, or mostly the same problem in different words?

[jinliyl]

@Alake-SwordFish
That's a really great idea—and I think it’s genuinely meaningful. The same user interacting across different models, devices, and sessions often faces inconsistent contexts and environments, which can lead to a fragmented, disjointed experience that feels erratic or “jarring.” A Presence Continuity Layer could serve as a unified user memory layer, enabling a more coherent and personalized experience regardless of where or how the user engages. It’s not just about remembering facts—it’s about preserving thefeeling of continuity, so the user always feels seen and understood, no matter the context.

[Alake-SwordFish]

Thanks, that’s really helpful to hear.
One thing I’ve been thinking about is whether this continuity layer should exist completely outside the model stack, almost like a persistent user substrate that different models can attach to.
Do you see something like that fitting into systems like ReMe, or would you imagine it more as a layer on top of agent memory systems?