Resources · n8n integration

n8n Calculator nodeConfigure Calculator in n8n.

An agent that writes a clean report can still get a multiplication wrong. The Calculator node hands it a real arithmetic engine instead of a guess. The n8n Calculator node has no panel to fill in and no account to link: it clips onto the Tool port of an agent and starts answering sums.

Try the Calculator node on n8n

Verified Trustpilot reviews · AI, automation & growth agency

Why automate

What does the n8n Calculator node actually do?

The Calculator node is a tool sub-node for n8n's AI nodes. A sub-node is a small node that never runs by itself: it clips onto a port of a root node, here the Tool port of an agent, and waits to be called. When the agent hits a number it cannot work out reliably, it sends the expression to the Calculator, gets the result back, and keeps reasoning with a figure it did not invent.

Language models are good at text and shaky at arithmetic. They produce a plausible-looking total rather than a computed one, and the error only shows up later, in a report someone forwards to a client. Handing the sum to a tool removes the guesswork: the agent still decides when to compute, the Calculator decides what the answer is.

Three shapes come back again and again. A marketing agent that pulls raw spend and conversions from several ad platforms, then needs a cost per lead before it writes a paragraph about it. A financial reporting agent that reads rows out of a database and has to total a month, compare it against the previous one, and phrase the difference. A data chatbot that answers questions in a chat window and gets asked for an average or a share of total. In all three, the reasoning belongs to the AI Agent and the number belongs here.

Adding it is close to free in effort. There is no panel to configure, so the whole decision is whether the agent should be allowed to do math at all. If the answer is yes, drag the node in, run one connection to the Tool port, and describe in the agent's prompt what it should compute. The same agent can keep its other tools next to it, including an AI Agent Tool when a second agent is exposed as a tool of the first.

There is a case for skipping it. If the calculation is fixed and known when you build the workflow, a monthly ratio that never changes shape, compute it in the workflow before the agent runs and pass the finished figure in. A tool call costs a round trip through the model, and a model that has the number already has nothing to decide. Calculator earns its place when the math depends on what the agent finds, not on what you planned.

The limits are worth knowing upfront. The node handles arithmetic, not data access: it cannot read a spreadsheet or query an API, so pair it with the tools that do. It also does nothing to the model's own judgment about which numbers matter, and that judgment comes from the model you branch onto the agent, whether that is OpenAI, Google Gemini or another provider. If you are still weighing platforms before building any of this, the n8n review covers the ground.

Connections

Where do you plug the Calculator node?

n8n's AI nodes follow a root node plus sub-nodes model: the root node receives the workflow items through its main input, and sub-nodes clip onto its ports, one type of sub-node per port. Calculator is a sub-node, so it has no main input and never runs on its own.

Output (what it plugs into)

  • Toolai_tool
01ToolThe node sends its output to the Tool port of a root node, the socket where an agent collects the tools it may call while it reasons. Once the wire is drawn, the agent decides by itself when a figure needs computing.

The node sends its output to the Tool port of a root node, the socket where an agent collects the tools it may call while it reasons. Once the wire is drawn, the agent decides by itself when a figure needs computing.

Key parameters

  • Required: the connection itself. With nothing on the other end, the node has no main input and nothing ever calls it.
  • AI Agent: the usual host, the root node that works through a task step by step and reaches for tools between steps.
  • AI Agent Tool: an agent exposed as a tool of another agent, so arithmetic stays reachable one level down.
  • MCP Server Trigger: publishes a workflow's tools to outside clients, Calculator included.
Use cases
a reporting agent that reads raw ad spend and has to state a cost per lead.
Parameters

Calculator

The Calculator node has 0 parameters. For each one: the node as you configure it in n8n, what the parameter changes, and our field notes.

Need help

Need help automating Calculator with n8n?

A person reads every message.

FAQ

Questions people ask about the Calculator node

01Is the Calculator node included in n8n?
Yes. It ships inside n8n's AI package, so there is nothing to install and nothing extra to pay on the n8n side, on n8n Cloud as well as on a self-hosted instance under the Community Edition and its Sustainable Use license. A workflow behaves the same either way, which matters if you prototype on Cloud and move to your own server later. The one cost attached to an agent workflow comes from elsewhere: the provider of the chat model you branch onto the agent bills its own API calls under its own terms. The Calculator itself adds nothing to that bill, since it computes locally rather than calling anyone.
02What do you need to set up before it works?
Nothing, and that is the point rather than a gap. The node has no credential, no authentication selector and no parameters at all, so there is no account to link and no panel to fill in. Drag it onto the canvas, connect it to the Tool port of a root node, and it works. Authentication lives one node over: the chat model sub-node branched onto the same agent carries the provider credential, which is the reusable stored login n8n keeps for you. The only real preparation is on the prompt side, telling the agent that it can compute rather than estimate.
03What are the limits of the n8n Calculator node?
It does arithmetic and nothing else. It will not fetch data, query an API, read a spreadsheet or reformat a result, so it belongs next to the tools that do those things rather than instead of them. Since it exposes no parameters, there is also no way to constrain it, no rounding rule to set and no unit to declare: whatever the agent sends is what gets computed. The node is at version 1, and pages describe the highest version in the catalog, so an older workflow may show something narrower. It also never runs alone, which means a disconnected Calculator on the canvas does nothing at all.
04What do you have to connect at minimum?
Three pieces, and the Calculator is the last of them. Start with a root node that actually reasons, an agent, because a sub-node has no main input and cannot be triggered by the workflow. Branch a chat model sub-node onto it, since that is what produces the reasoning and carries the provider credential. Then run the Calculator into the Tool port. If you are building your first one, keep it to those three and add memory or extra tools once the agent behaves. A frequent beginner trap is wiring Calculator straight into the workflow line and wondering why it never fires.
05n8n or Make for an agent that needs math?
It depends on where you want the workflow to live and how you want to pay for it. Make is hosted by Make, with no self-hosting option, and bills per operation, so a chatty agent that calls a tool several times per run is billed several times. n8n runs on n8n Cloud or on your own server, which matters when the data going through the agent cannot leave your infrastructure. The visual logic differs too: n8n branches and loops as a graph, Make walks a scenario in sequence. Both get the job done, so pick on hosting, data control and cost shape rather than on features.
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