Resources · n8n integration

n8n Simple Memory nodeConfigure Simple Memory in n8n.

Chat history is what turns a one-shot answer into a conversation. The n8n Simple Memory node keeps that history inside n8n itself, with 2 parameters to think about: Session Key and Context Window Length. It clips onto the Memory port of a root node, and there is no account to link.

Try the Simple Memory node on n8nSee the 2 parameters

Verified Trustpilot reviews · AI, automation & growth agency

Why automate

What does the n8n Simple Memory node actually keep?

Simple Memory holds the last turns of a chat in n8n's own memory and hands them back to a root node on every run. Nothing is stored outside n8n, and the node description says it plainly: stores in n8n memory, so no credentials required. It is a sub-node, a small node you attach to a port of a bigger one instead of dropping it in the main flow, so it never executes on its own.

Start with the case everybody hits first: a chat assistant that has to remember what was said two messages ago. A Chat Trigger feeds an AI Agent, a chat model such as one of the OpenAI sub-nodes does the talking, and Simple Memory sits on the Memory port so each new question arrives with the previous exchanges attached. The Session Key is picked up automatically from the On Chat Message trigger, which is why this setup works before you touch a single parameter.

Second case: one workflow, several conversations that must not bleed into each other. Every Simple Memory node in a workflow points at the same memory instance by default, so two agents left on the default settings end up reading each other's history. Giving each node its own Session Key is what separates them. The same care applies before any destructive step, like the override all messages operation in the Chat Memory Manager node, which rewrites whatever is in there.

Third case: the conversation that grows too long. Context Window Length decides how many past interactions the model receives as context. Keep it small and the model sees a short recent window; raise it and it sees more of the thread, and every one of those past turns travels with the request to the provider. That matters because the provider behind the chat model, Anthropic or another one, bills its own API calls under its own terms while n8n adds nothing on top.

There is one place where this node is the wrong pick, and the official docs are blunt about it: an n8n instance running in queue mode. In that setup n8n cannot guarantee that every call to Simple Memory lands on the same worker, so the node does not work in an active production workflow. Check how your instance runs before you build a customer-facing bot on it.

The rest of the limits are easy to live with. Memory is a thing root nodes use: an AI Agent or an AI Agent Tool can take a memory sub-node, chains cannot. And if you are still deciding which platform to build this on, the n8n review goes through the trade-offs in detail.

Connections

Where does this sub-node plug in?

n8n's AI nodes work as a root node plus sub-nodes: the root node receives the workflow items on its main input, and each sub-node clips onto one of its ports, one type per port.

Output (what it plugs into)

  • Memoryai_memory
01MemorySimple Memory has no main input at all. Its only connection is the ai_memory port of a root node, and that link is what makes it run: on its own it does nothing.

Simple Memory has no main input at all. Its only connection is the ai_memory port of a root node, and that link is what makes it run: on its own it does nothing.

Key parameters

  • Required: the node never executes alone, so it has to be attached to a root node that accepts a memory sub-node.
  • AI Agent and AI Agent Tool: the two agent nodes that read the stored history before calling their model.
  • Chat Memory Manager and Chat Trigger: the other root nodes in the catalog that accept this port.
Use cases
a support bot keeps its thread on an AI Agent, while a Chat Memory Manager on the same memory lets you inspect or reset that thread from another branch.
Parameters

Which 2 parameters are worth your time?

The Simple Memory node has 2 parameters. For each one: the node as you configure it in n8n, what the parameter changes, and our field notes.

01

Session Key

sessionKey

What you see in n8n

Notes & use cases

This is the key used to store the memory in the workflow data. One key means one conversation: the same value brings back the same history, a different value starts a clean thread.

Key parameters

  • Session Key: a string, usually left alone since the sessionId comes from the On Chat Message trigger; set it by hand with {{ $json.sessionId }} when the chat starts elsewhere.
Use cases
with two agents in one workflow, give each its own key so they stop sharing a thread. While testing, a static key like my_test_session is fine, but swap it for real session handling before the workflow goes live. An error reading No sessionId means the trigger output has no session to read.
02

Context Window Length

contextWindowLength

What you see in n8n

Notes & use cases

Here you set how many past interactions the model receives as context. It is the dial between an agent that remembers everything and one that glances at the last few turns.

Key parameters

  • Context Window Length: a number, counted in previous interactions, not in characters or in tokens, the units a model bills and measures text in.
Use cases
a short FAQ bot rarely looks far back, so a low value keeps each request light. A booking assistant that collects a date, a name, then a confirmation still needs to see the date at the third message. Raise it until the agent stops asking twice, no further: every remembered turn is sent again on the next call.
Need help

Need help automating Simple Memory with n8n?

A person reads every message.

FAQ

Simple Memory in n8n, the questions that come next

01Is the n8n Simple Memory node free and included in n8n?
Yes. It ships with n8n as part of the AI package, @n8n/n8n-nodes-langchain, so there is nothing to install and nothing extra to pay on the n8n side. That holds both on n8n Cloud, the hosted offer run by n8n, and on a self-hosted instance under the Community Edition and its Sustainable Use license. A workflow behaves the same way in either place, which means you can build it locally with Docker or npm and move it to Cloud without rewriting the memory part. The only cost attached to an AI workflow comes from the model provider behind the chat model you plug into the same agent, billed by that provider under its own terms.
02What credentials does Simple Memory need to work?
None, and that is the point of this node. Its own description is short: it stores in n8n memory, so no credentials required. There is no Authentication selector, no API key field, no account to connect, which also means there is nothing to rotate or to break later. The credential lives one node further, on the chat model sub-node you attach to the same agent, and that one does carry the provider key. So the setup order is simple: connect the model with its credential, then drop Simple Memory onto the Memory port of the root node and give it a Session Key if the trigger does not supply one.
03What are the limits of Simple Memory in n8n?
The big one is queue mode. If your n8n instance runs in that mode, the node does not work in an active production workflow, because n8n cannot guarantee that every call to Simple Memory reaches the same worker. The second limit is sharing: add several Simple Memory nodes to one workflow and they all point at the same memory instance unless you set different session IDs. Be careful around destructive steps such as the override all messages operation in the Chat Memory Manager node. Last, memory belongs to agents: AI agent nodes can use it, AI chains cannot. This page describes version 1 of the node, an older workflow may show fewer options.
04What do you need to plug in for an agent with memory?
Two sub-nodes at minimum on the root node. A chat model, which is the sub-node that actually generates the answer and holds the provider credential, and a memory sub-node if you want the conversation to continue across messages. Simple Memory is the one to start with, because it needs no service and no account. Wire it to the Memory port of an AI Agent, keep the Session Key coming from the On Chat Message trigger, and leave Context Window Length at its default until the agent visibly forgets something. Add the Chat Memory Manager later, when you need to read or clear that history from another branch of the workflow.
05n8n or Make for an AI chatbot with memory?
It comes down to four questions, and neither tool is the wrong answer. Hosting: n8n runs on your own servers with Docker or npm, or on n8n Cloud, while Make is hosted only. Data control: memory kept inside your own n8n instance never leaves it, which weighs a lot when the conversation carries customer details. Cost model: Make bills per operation, so a chatty agent that fires many calls per conversation adds up differently. Visual logic: both are canvas tools, the shapes just differ, n8n leaning on sub-nodes plugged into ports. Try a single conversation flow on each and see which one you can debug faster.
Hack'celeration Lab

Get our weekly integration tips.

No spam. Unsubscribe anytime.

Same platform

Other docs you might need on this stack.

Browse all integrations