Skip to main content
0 reactions

n8n Webhook Test URL vs Production URL: What’s the Difference?

Card titled Test URL vs Production URL, labelled n8n webhooks, comparing two cards: Test URL with /webhook-test/…, One call at a time and Data on the canvas, and Production with /webhook/…, Once published and In Executions.

Every n8n Webhook node has two addresses, a Test URL and a Production URL. They look almost the same, but they answer at different times, and calling the wrong one is the usual reason a webhook replies with a 404.

This post shows both with the smallest possible example, a Ping workflow that answers a GET request at /ping, called with curl from Terminal. By the end you know which address to use while you build, which one to give other apps, and where each run shows up.

Cost: free. Everything here runs on your own n8n Community Edition.

Every command, screen and output on this page is real, run with n8n 2 in Docker on macOS.

Test URL vs Production URL at a Glance

Test URLProduction URL
Address/webhook-test/ plus the path/webhook/ plus the path
When it answersOne call, after you click Listen for test event or Execute workflowEvery call, while the workflow is published
How long it waits120 secondsUntil you unpublish the workflow
Where the run showsOn the canvas, in the node’s OUTPUTIn the Executions tab
Use it forBuilding and testingOther apps and real traffic

The listening time comes from n8n’s webhook common issues page (opens in a new tab); everything else in the table is shown live below.

Step 1: Create a Ping Webhook

Create a new workflow named Ping, click Add first step and choose On webhook call. In the Webhook node, keep HTTP Method at GET and Respond at Immediately, and set Path to ping. At the top, Webhook URLs shows the Test URL:

The Webhook URLs section of an n8n Webhook node with Test URL selected, showing GET http://localhost:5678/webhook-test/ping.
The Test URL contains webhook-test.

Step 2: Call the Test URL

Open Terminal (press ⌘ Space, type Terminal and press Return) and call the Test URL with curl. Given just an address, curl sends a GET request:

TerminalBash
curl http://localhost:5678/webhook-test/ping

n8n answers with a 404 error, because nothing is listening yet:

OutputPlain text
{"code":404,"message":"The requested webhook \"ping\" is not registered.","hint":"Click the 'Execute workflow' button on the canvas, then try again. (In test mode, the webhook only works for one call after you click this button)"}

Click Listen for test event in the Webhook node. The node now waits for one call:

The Webhook node listening for a test event, asking for a GET request to http://localhost:5678/webhook-test/ping, with a Stop Listening button.
Listen for test event waits for one call.

Run the same curl command again. This time n8n starts the workflow:

OutputPlain text
{"message":"Workflow was started"}

The request appears in the node’s OUTPUT, with executionMode set to test:

The Webhook node output after a test call: headers from curl, empty params, query and body, webhookUrl http://localhost:5678/webhook-test/ping and executionMode test.
A test call appears on the canvas.

Run curl a third time and the same 404 is back: the Test URL takes one call each time you click Listen for test event or Execute workflow.

Step 3: Call the Production URL

Click Production URL in the Webhook node. The address is the same without -test:

The Webhook URLs section with Production URL selected, showing GET http://localhost:5678/webhook/ping.
The Production URL drops -test.

Call it before publishing the workflow:

TerminalBash
curl http://localhost:5678/webhook/ping
OutputPlain text
{"code":404,"message":"The requested webhook \"GET ping\" is not registered.","hint":"The workflow must be active for a production URL to run successfully. You can activate the workflow using the toggle in the top-right of the editor. Note that unlike test URL calls, production URL calls aren't shown on the canvas (only in the executions list)"}

The production address only answers while the workflow is published. Close the panel, publish the workflow and run the same command again:

OutputPlain text
{"message":"Workflow was started"}

Nothing changes on the canvas: production runs aren’t shown there. Open the Executions tab: the production call is at the top, and the run with the flask icon is your test from Step 2.

The Executions tab of the Ping workflow with two successful runs: the production call at 19:36:34 without an icon and a test call at 19:35:58 with a flask icon.
Production runs show here; the flask marks a test.

Which URL Should You Use?

Use the Test URL while you build: you see each request on the canvas and can map its fields into the next nodes. When the workflow is ready, publish it and give other apps the Production URL. Both addresses here start with localhost, so they only work on the computer that runs n8n.

Clean Up

Unpublish the Ping workflow, then archive and delete it from its ⋯ menu. Deleting it can’t be undone.

Common Mistakes

MistakeWhat happensDo this instead
Calling the Test URL without clicking Listen for test event or Execute workflowA 404 error: the webhook isn’t registeredClick one of them, then call within 120 seconds
Giving another app the Test URLOnly the call right after you click worksPublish the workflow and give apps the Production URL
Calling the Production URL before publishingA 404 error with a hint that the workflow must be activePublish the workflow

Troubleshooting

ProblemCauseFix
The requested webhook "ping" is not registered. with the hint to click Execute workflowYou called the Test URL while nothing was listeningClick Listen for test event or Execute workflow, then call again
The requested webhook "GET ping" is not registered. with the hint that the workflow must be activeYou called the Production URL of a workflow that isn’t publishedPublish the workflow
A production call doesn’t show on the canvasProduction runs only show in the execution listOpen the Executions tab

FAQ

Why does the Test URL stop answering after one call?

It’s made for building: each click registers it for one call, so you can look at that request on the canvas. n8n’s own hint says it: in test mode, the webhook only works for one call after you click the button.

Do test calls show up in the Executions tab?

Yes. In Step 3 the test call from Step 2 is in the list too, marked with a flask icon.

Can I change the webhook-test and webhook parts of the address?

On self-hosted n8n, yes: the N8N_ENDPOINT_WEBHOOK_TEST and N8N_ENDPOINT_WEBHOOK settings set them (n8n’s endpoint settings (opens in a new tab)).

Conclusion

The Test URL is for you: it answers one call after you click, and shows the request on the canvas. The Production URL is for other apps: it answers every call while the workflow is published, and its runs go to the Executions tab. A 404 almost always means the address you called isn’t listening yet.

Additional Resources

n8n and other names mentioned are trademarks of their respective owners. JD Bots isn’t affiliated with or endorsed by n8n GmbH.

0 comments