The Continuity Layer
Definition

What Is a Continuity Layer for AI?

thecontinuitylayer.com/what-is-a-continuity-layer

The canonical definition this project uses, and the anchor other pages cite. It is deliberately narrow — a continuity layer is a specific part of an AI system, not a synonym for memory.

Canonical definition
A continuity layer is the part of an AI system that carries forward governed state across sessions, tools, and models: what is currently true, what changed, why it changed, and which evidence supports it. It sits above memory and retrieval, and below agent execution, turning stored history into usable, auditable continuity.
The Continuity Layer
01

Where it sits

A continuity layer is defined by its position in the stack, not by any one feature.

It sits above the memory and retrieval infrastructure that stores and fetches facts, and below the agents that execute. It consumes what memory holds, governs what counts as current, and — by definition — hands every surface, every model, every tool, the same governed state. Because the layer improves whenever the memory beneath it improves, it is a layer, not a feature.

02

What it carries forward

Governed state is four things held together, each auditable.

What is currently true — the present state of the project, not a transcript of everything that happened.
What changed — the moves that took the project from its old state to its current one.
Why it changed — the reason attached to each change, so a decision can be re-examined, not just observed.
Which evidence supports it — the provenance of each fact — where it came from, so any claim can show its work.
03

What it is not

The distinctions that keep the term precise.

Not memory alone

memory stores the past; continuity governs what of it is still true.

Not retrieval

retrieval guesses what might be relevant to a query; it doesn't decide what is current.

Not long context

a bigger window extends one session; it doesn't carry the session forward.

For the fuller structural version of these distinctions, see the Primer’s Memory Is Not Continuity, and the side-by-side breakdown in AI Memory vs AI Continuity.