Connect a service to n8n
Updated 2026-11-13
How to use a published service in n8n workflows: install the SpreadAPI node, connect it with the service URL, pick inputs, and use the node as a tool for n8n AI agents.
What you get
The SpreadAPI node for n8n runs a calculation of your service as one step of a workflow. You paste one address into n8n, choose the inputs from a list, and the results come back as fields for the next node. You do not build HTTP requests by hand.
Before you start
- The service must be published. Until then the setup page says "Publish the service before you connect it to n8n."
- If the service requires a token, have a token ready. A token is shown only once, when it is created; if you no longer have it, create a new one under API → Access (Zugriff). See Tokens and access.
Get the service URL
- Open your service and open API → Automation → n8n (Automatisierung).
- Under Service URL for n8n (Service-URL für n8n), click Copy (Kopieren).
The address looks like this:
https://spreadapi.io/d/YOUR_SERVICE_ID
If the service requires a token, the page shows the note This service needs a token. Add your token to the copied address before you paste it into n8n:
https://spreadapi.io/d/YOUR_SERVICE_ID?token=YOUR_TOKEN_HERE
Set up in n8n
The names of n8n's own menus are given as n8n shows them.
- Install the SpreadAPI node (once). In n8n open *Settings → Community Nodes → Install* and enter
n8n-nodes-spreadapi. On n8n Cloud, search for "SpreadAPI" in the node panel. - Add the SpreadAPI node to your workflow. Click + in the workflow editor, search for "SpreadAPI" and add the node.
- Create the credential with the service URL. In the node choose *Create new credential* and paste the address into Service URL. n8n tests the connection when you save.
- Pick your inputs and run. Click *Add Input*: the service's inputs are offered in a list. Set the values or map them from earlier nodes, then execute. The results come back as fields for the next node.
Good to know
- One calculation per item. The node runs once for every incoming item; 100 items are 100 calculations. Each is one call of your service.
- Percentages are decimals. Send 5% as
0.05. The input list marks percentage fields. See Percentages and formats. - One credential, one service. The service URL identifies one service. For a second service, create a second credential with that service's URL.
- After changes to the service. When you add or rename inputs and publish again, open the node and check the input list, so that the workflow sends the names the service now expects.
AI agents in n8n
The SpreadAPI node also works as a tool for n8n AI agents and appears in their tool list. The agent then calculates with your workbook instead of estimating numbers. Describe the service well, so the agent knows when to use it and what each input means: see AI → Definition.
Limits
- Calls from n8n count like every other call: toward the calls of your plan, and at most 1,000 calls per minute per service. See Call limits.
- When a value is refused, the service answers with a message that names the input and what is wrong. See Input errors.
- The node calls a service; it does not start a workflow when something changes in SpreadAPI.
Related topics
- Zapier and Power Automate
- Other tools over HTTP — also the way to call a service from n8n's HTTP Request node
- The discovery address
For developers and agents
| Item | Value |
|---|---|
| Package to install | n8n-nodes-spreadapi |
| Credential field | Service URL |
| Value for a public service | https://spreadapi.io/d/{serviceId} |
| Value for a protected service | https://spreadapi.io/d/{serviceId}?token={token} |
| Where to copy it | API → Automation → n8n in the service |
The service URL is the service's discovery address. A GET on it returns the service's description with its inputs (name, title, type, required, allowed values) and outputs; this is what fills the input list in n8n. It answers 404 with "Service not found or not published" for a service that is not published, and 401 when a required token is missing or wrong.
Input names are not case-sensitive. Send numbers as numbers, yes/no values as true or false, and percentages as decimals.
FAQ
How do I connect my service to n8n?
Install the community node n8n-nodes-spreadapi, add the SpreadAPI node to your workflow and create its credential with the address from API → Automation → n8n. Then pick your inputs with *Add Input* and execute.
Where do I find the service URL for n8n?
In your service under API → Automation → n8n, in the field Service URL for n8n. It has the form https://spreadapi.io/d/YOUR_SERVICE_ID.
My service needs a token. Where does it go in n8n?
Add it to the service URL before you paste it into the credential: https://spreadapi.io/d/YOUR_SERVICE_ID?token=YOUR_TOKEN_HERE. If you no longer have the token, create a new one under API → Access.
Why does the n8n page say I have to publish first?
Automation tools call the published version of a service. Click Publish in the service, then open API → Automation → n8n again.
How do I install the SpreadAPI node in n8n?
In n8n open *Settings → Community Nodes → Install* and enter n8n-nodes-spreadapi. On n8n Cloud, search for "SpreadAPI" in the node panel.
Why does my workflow use so many calls?
The node runs once for every incoming item, so 100 items are 100 calculations. Reduce or filter the items before the SpreadAPI node if you need fewer calls.
How do I send a percentage from n8n?
As a decimal: 0.05 for 5%. The input list in the node marks percentage fields.
Can an n8n AI agent use my service?
Yes. The SpreadAPI node works as a tool for n8n AI agents and appears in their tool list. A clear description under AI → Definition helps the agent use it correctly.
Can I use several services in one workflow?
Yes. Create one credential per service, each with that service's URL, and add one SpreadAPI node per calculation.
As an agent, what do I need from the user to set up the n8n credential?
The service URL https://spreadapi.io/d/{serviceId}, and for a protected service the token appended as ?token={token}. The service must be published; a GET on the address returns the inputs and outputs to use.