---
title: "What is agent memory, and where does it fit?"
description: "The three kinds of agent memory, what each one stores, and how they sit alongside the files, databases and graphs an AI stack already runs on."
url: https://surrealdb.com/explore/what-is-agent-memory
---

# Agents forget everything, every single time.

[Skip to content](https://surrealdb.com/explore/what-is-agent-memory#kinds)

Agent memory

Every session starts from nothing. The agent re-fetches the same context, re-reads the same documents and re-derives the same conclusion a colleague reached last Tuesday. That is not a model problem, and a longer context window does not fix it. It is a missing layer.

Files and databases hold **current state**. Agent memory holds **everything that was ever true**.

The graph says what is true now. Memory says why, and what you already concluded.

**3** kinds of memory: facts, episodes, preferences

**1** engine underneath

**~74%** fewer tokens, illustrative

**0** cold starts

[Explore SurrealDB Agent Memory](https://surrealdb.com/agent-memory)

Part one

## What agent memory actually is

Ask three vendors what memory means and you get three answers. Underneath the vocabulary there are only three things worth storing, plus the thing that gets them back. Each card below opens onto a real memory captured by a working agent, not an invented example.

> Search asks "what is most similar?". Recall also asks "what is recent, what matters, and what is related?" *SurrealDB for AI engineers, lesson 06*

Part two

## Where it fits in what you already run

Two different questions, and they get different answers. The systems your data already lives in stay exactly where they are: a context layer reads from them. The stores you added to serve the agents are the ones worth consolidating. Tick both rows and the diagram redraws around you. Not everything collapses, and anyone who tells you otherwise is selling.

Architecture diagram: your sources, the systems you run today, one context layer with agent memory beside it, and the agents that read from it

Scroll the diagram sideways to see the whole flow

> The composition happens in the context window, and the context window has no access-control model. It is a text buffer. The architecture had no place to say no. *SurrealDB for AI engineers, lesson 08*

Part three

## One question, two architectures

The same request, answered twice. On the left an agent with no memory. On the right the same agent with a memory layer underneath it. Step through and watch where the tokens go.

Step forward

Reset

### No memory

Re-derives the answer from source every time

**0** tokens

### With agent memory

Recalls what was already concluded, then writes back

**0** tokens

**These figures are illustrative, not a benchmark.** They come from an internal worked example of a single support-style question, and nothing here has been measured against your workload. The shape of the saving is the durable part: the cost of re-derivation scales with how much of your context is genuinely re-derivable. Your own numbers depend on prompt size, how verbose your tools are, and how often the same question gets asked twice. Ask for a measurement against your traffic before you put a figure in a business case.

Part four

## Why this is a budget line, not a feature request

The interesting thing about a memory layer is not what it lets a clever engineer do. It is what it stops everyone else from having to redo.

> Knowledge becomes a write-time cost, not a read-time tax. *Spectron Code, product brief*

### It raises the floor

It does not make your best engineer better. It makes every session start where the last one finished, which is where most of the waste actually lives.

### It compounds three ways

With team size, with the number of projects, and with time. A knowledge layer is worth more in month twelve than in month one, which is the opposite of most tooling.

### It has a place to say no

Permissions travel with the data and the filter runs before ranking. The same question returns different rows to different people, in the database, not in a prompt.

[See SurrealDB Agent Memory](https://surrealdb.com/agent-memory)

## Explore more

- [**Financial services ontology** Eight source systems land in one SurrealDB graph, with the SurrealQL behind every entity and relation.](https://surrealdb.com/explore/financial-services-ontology)
- [**Company brain** One pricing decision is followed through ingest, per-person permissions and time.](https://surrealdb.com/explore/company-brain)