N8n-hosted MCP client and server setup throws errors

Hi there :slight_smile:

I am building a MCP client and server setup hosted by n8n. For each (client and server) I created a workflow:

Server workflow:

Client workflow:

Before sending a chat message to the AI agent I need to click the “Execute workflow” message in the server. After doing that it looks like the following:

Now I can send a chat message and the server works fine.

When I am sending the second request the server is again in the state of the initial screenshot (the “Execute workflow” button is available again). When I don’t click it before sending another request I am receiving an error message as follows:

That means, every time before I send a new request to the server I need to enable it by clicking the “Execute workflow” button.

I read several threads but all of them are related to self-hosted n8n instances - but in my case I let n8n host it so this might not apply. Do you have any idea on how to solve this?

Thanks,

Peter

Ah, my bad. When I am using the production URL it works fine. Only as long as I use the testing URL I need to start the workflow first.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.