Step 2: Connect the Web Scraping Source

Scrape the provided URL and map the input values into a clean payload for the AI node.

1. Add the Scrape Web Page node and connect it after Form Submit Trigger.
2. Set Resource to extraction and Operation to scrape.
3. Set URL to ={{ (() => {\n const url = ($json['Page URL'] || '').trim();\n return url && (url.startsWith('http://') || url.startsWith('https://')) ? url : 'http://' + url;\n})() }}.
4. Credential Required: Connect your airtopApi credentials.
5. Add Map Input Fields and set assignments:
   • URL → ={{ $('Form Submit Trigger').item.json['Page URL'] }}
   • Instructions → ={{ $('Form Submit Trigger').item.json.Instructions }}
   • Post → = -
   • =Webpage_content → ={{ $json.data.modelResponse.scrapedContent.text }}

Step 3: Set Up the AI Drafting Pipeline

Generate a draft post based on scraped content and user instructions, then assemble the output.

1. Add Compose Post Draft and connect it after Map Input Fields.
2. Set Text to =Write a helpful educational, and engaging LinkedIn posts based on a webpage scraped content below or the Post text below and according to the instructions provided by the user.\n\nInstructions: {{ $json.Instructions }}\n\n{{ $json.Post }}\n\n{{ $json.Webpage_content }}.
3. Set Resource to assistant and select the configured assistant ID.
4. Credential Required: Connect your openAiApi credentials.
5. Add Assemble Draft Output and set assignments:
   • URL → ={{ $('Map Input Fields').item.json.URL }}
   • Instructions → = -
   • Post → ={{ $json.output }}

Step 4: Configure Slack Review and Feedback Classification

Send the draft to Slack for approval, then classify the response as Approved, Edit, or Declined.

1. Add Slack Review Request and connect it after Assemble Draft Output.
2. Set Operation to sendAndWait and Response Type to freeText.
3. Set Message to =----------------\n\n{{ $json.Post }}\n\n-----------------\nBased on: {{ $json.URL }}.
4. Credential Required: Connect your slackOAuth2Api credentials.
5. Add Evaluate Text Feedback and set Input Text to ={{ $json.data.text }}.
6. Verify the Categories include Edit, Approved, and Declined with the provided descriptions.
7. Ensure Language Model Chat is connected to Evaluate Text Feedback via the AI language model connection and set Model to gpt-4.1.
8. Credential Required: Connect your openAiApi credentials in Language Model Chat.

Step 5: Configure Revision Loop and Final Publishing

Handle revision requests and publish the final post to Slack when approved.

1. Add Apply Revision Inputs and connect it from Evaluate Text Feedback for edit requests.
2. Set assignments in Apply Revision Inputs:
   • URL → ={{ $('Map Input Fields').item.json.URL }}
   • Instructions → ={{ $json.data.text }}
   • Post → ={{ $('Assemble Draft Output').item.json.Post }}
3. Connect Apply Revision Inputs back to Compose Post Draft to regenerate the draft with revisions.
4. Add Publish Final Post and connect it from Evaluate Text Feedback for approved responses.
5. Set Text in Publish Final Post to =---------------- \nFINAL POST:\n---------------- \n\n{{ $('Assemble Draft Output').item.json.Post }}\n.
6. Credential Required: Connect your slackOAuth2Api credentials.

Step 6: Test and Activate Your Workflow

Validate the full approval loop and enable the workflow for production use.

1. Click Execute Workflow and submit the form in Form Submit Trigger with a valid URL and instructions.
2. Verify Scrape Web Page returns content and Compose Post Draft generates a draft in Assemble Draft Output.
3. Check Slack for the draft message from Slack Review Request and reply with “Approved” or a revision request.
4. Confirm a revision reply loops through Apply Revision Inputs and updates the draft; an approval sends the final message via Publish Final Post.
5. Once validated, toggle the workflow to Active to enable production use.