Hi @mikemoz, thanks for reaching out! I just pulled down the same workflow to test. It looks like you were testing using your n8n cloud account so I did as well, version 1.42.1 and following the documentation for the embedded chat option. I’m assuming you used a simple html page in your browser to test based on the n8n.html filepath in your editor.
The error that you shared in the n8n AI error popup indicates that the format of the AI node output could not be processed and is unrelated to the CORS configuration.
I was not able to replicate your console output. The first error that you received seems to indicate that you did not specify a valid value for the Allowed Origins option.
When I set a restrictive value for the Allowed Origins option I could see a corresponding error in my dev tools’ network panel indicating that there was an error in the workflow. The more generous catchall resulted in no error in my network panel.
Could you share any error messages that you see in your network panel when testing?