How to manually check output of a node before passing it on to the next node

Questions
1

Describe the problem/error/question

I’m using a ChatGPT node to summarise some text. Is there a way to manually check this summary before posting it to X, for example?

What is the error message (if any)?

N/A

Please share your workflow

N/A

Share the output returned by the last node

N/A

2

It looks like your topic is missing some important information. Could you provide the following if applicable.

  • n8n version:
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system:
3

Hello @gimi !

Welcome to out n8n community! It is great to see you here.

I have read your post and if I understood it right, you have a workflow that sends text to chatGPT to summarise it for you, which it does. However you would like to have a manually check on it before going ahead on the target system(X for example). I am assuming that by manually you mean having a human(a person) to read it and then to give the go ahead or an no-go decision. Would this be correct?

With the assumption of the above, I will also assume it would be an asynchronous task. meaning, the chatGPT summary could wait till someone would be available to assess it.

With that said, my first idea(maybe it is too simple and others may come up with more interesting options) would be to have 2 workflows:

1- One to have the text capture and summarised and then they summary would be parked somewhere(email, spreadsheet, target system, etc) waiting for an evaluation. Once the evaluation is done then it would go to the second workflow to direct the approved summarised text to the target system or reject it and perhaps execute any action on it (send it back to ChatGPT with more insights of what is missing etc).

Let me know if my initial overview of the use-case gives you any positivie insight or if you are expecting something else/different as example for a solution in a single workflow which then I would love to hear from others.

i hope this is a good start of the brainstorm.

Cheers,

Flavio

4

You will have to be more specific about what you want. From where do you want to check? Are you talking about checking inside the workflow? You can always see the output of the GPT node by clicking on it.
If you are talking about pausing an automatic workflow to wait for human input, you could, for example, send yourself an email, then when you want to confirm, use a clickable link in the email which triggers a Webhook which then sends off the X post.

5

Hi Flavio_Orfano.

This is a great idea.

Thank you. :+1:

6

Hi JayF.

Thanks for your answer.

Could you explain how the clickable link idea works?

Is all the data for the X post and the webhook url contained within the clickable link some how?

Thanks again

7

Hi gimi
So it’s just a matter of clicking on your Webhook trigger and copying the Production URL. When anyone accesses that URL, it will start the corresponding workflow :sunglasses: