What this workflow does
This workflow fetches content from a web page and uses AI to extract and analyze information automatically.
It saves people from doing long manual work by getting answers like page summary, authors, posts, and even a funny comment.
The result is organized data gathered quickly with little effort.
Who should use this workflow
This is for anyone who needs fast and accurate content details from web pages without coding skills.
It suits digital marketers, content planners, or researchers who want to spend less time extracting info and more on using it.
Tools and services used
- n8n workflow platform: To build and run the automation.
- Anthropic Claude 3.7 Sonnet API: Provides AI model for all natural language tasks.
- HTTP Request node: To get live page content.
- Markdown node: Converts HTML to markdown for easier AI reading.
- Memory nodes: Manage session context during AI prompt chaining.
- Webhook node: Accepts external prompt inputs for dynamic use.
Input → Processing → Output
Inputs
The input is a web page URL where content lives.
And optionally, prompt arrays via webhook for parallel AI calls.
Processing steps
- Get raw HTML from the webpage.
- Change raw HTML to markdown text.
- Clear old conversation memory.
- Create prompt questions about the page in a set node.
- Change prompt questions into an array form for looping.
- Split prompts into single items for one-by-one AI processing.
- Run several AI calls one at a time asking to summarize, list authors, list posts, and make humor.
- Also run all prompts in parallel via webhook and multiple AI calls for speed.
- Combine all AI answers with original prompt details into one final output.
- Store conversation memory for better context in future runs.
Outputs
The final output is a combined JSON which has all the AI answers to prompts along with the original webpage text summary.
This can be used directly to help content teams or plugged into other systems for deeper use.
Beginner step-by-step: How to build this in n8n
Download and Import
- Download the workflow file using the Download button on this page.
- Open the n8n editor and choose “Import from File” to bring in the workflow.
Configure API Keys and Settings
- Add your Anthropic API Key in the credentials section of each Anthropic Chat Model node.
- Check the HTTP Request node for the URL and change if you want another web page.
- Update session keys or IDs if your memory nodes or webhooks require specific keys.
Test and Activate
- Run the manualTrigger node by clicking “Test workflow” to make sure it works.
- Verify the combined output contains expected content like summaries and lists.
- Activate the workflow in n8n to set it live for regular or webhook-triggered use.
For users deploying in production, consider running self-host n8n for more control and performance.
Customization ideas
- Change URL in the HTTP Request node to analyze different web pages.
- Edit prompts in the “Initial prompts” node to ask new questions, like SEO tips or sentiment.
- Adjust Anthropic model settings like temperature for more creative or straightforward answers.
- Add new LLM Chain nodes for more detailed analysis or extra data extraction.
- Improve memory use in agent nodes to keep longer context for better AI understanding.
Common edge cases and failures
- If the webhook URL is not working, confirm the environment variable
$env.WEBHOOK_URLis set correctly in n8n settings. - Empty AI outputs can happen if prompt expressions or inputs are missing, check the data flows carefully.
- Memory problems usually relate to wrong session keys or too small context window—adjust these settings.
- Errors fetching the web page may come from wrong URLs or connectivity issues; verify URLs and network access.
Summary
✓ Saves hours of manual webpage research.
✓ Gives clear, structured info about page content and authors.
✓ Runs work in sequential and parallel AI calls for quality and speed.
✓ Uses memory nodes to keep good conversation context.
✓ Easy to import and configure in n8n with simple steps.
{
"system_prompt": "You will help analyze web page content.",
"step1": "What is on this page?",
"step2": "List all authors on this page.",
"step3": "List all posts on this page.",
"step4": "Write a humorous comment about the content."
}
{
"webhook_test_payload": {
"prompts": [
"What is on this page?",
"List all authors on this page.",
"List all posts on this page.",
"Write a humorous comment about the content."
]
}
}