Scrape and Summarize Webpages with AI in n8n: Turn Long Articles into Key Takeaways

Collect links from any index page, fetch every article and summarize each one with GPT-4o in a single n8n workflow.

3 min read

Research is mostly reading. Whether you track industry blogs, competitor news or long-form essays, the hard part is getting through it all. This use case shows a simple n8n workflow that scrapes a list of article links from an index page, fetches each article and produces an AI summary with GPT-4o. It's one of the best starter templates for combining scraping and AI in n8n.

Why Automate Web Research Summaries

  • Too much to read: dozens of long articles every week.

  • Copy-paste fatigue: pasting pages into ChatGPT one by one.

  • Inconsistent notes: summaries in different formats and places.

  • No repeatable process: research restarts from scratch every time.

Want summaries delivered to Slack, Notion or email on a schedule? Get a custom n8n workflow from Vyrade.

Download & Inspect the JSON Workflow

If you prefer starting from a ready-made workflow, subscribers can download the preconfigured n8n workflow JSON below and import it directly into their workspace.

The Automated Scrape-and-Summarize Workflow

Step 1: Start the run

A Manual Trigger starts the workflow. Swap it for a Schedule Trigger for daily digests.

Step 2: Fetch the index page

An HTTP Request node downloads the page that lists the articles (the template uses Paul Graham's essays page as an example).

Step 3: Extract links

The HTML node uses CSS selectors to pull every article link. Split Out turns them into individual items and Limit caps how many you process per run.

Step 4: Fetch each article

A second HTTP Request fetches each page and the HTML node extracts the main text and title.

Step 5: Summarize with GPT-4o

A Summarization Chain with a Recursive Character Text Splitter handles long pages by chunking, then OpenAI GPT-4o writes the summary.

Step 6: Merge results

A Merge and Set node combine title, URL and summary into one clean record per article.

What You Need Before You Start

  • n8n version 1.50.0 or later (n8n on Vyrade).

  • An OpenAI API key.

  • The URL of an index page and the CSS selector for its links. The n8n HTML node docs explain selectors.

How to Customize the Workflow

  • Save the output: write summaries to Google Sheets, Notion or Airtable.

  • Send a digest: email or Slack a daily roundup.

  • Change the prompt: ask for bullet takeaways, quotes, SEO keywords or tweet drafts.

  • Monitor RSS instead: replace link extraction with an RSS Feed Read node.

  • Dynamic sites: use the agent-based scraper from our AI agent web scraping use case.

Who Benefits From This Workflow

Content marketers build research briefs faster. Analysts track competitor blogs and news. Founders keep up with industry thinking in minutes. Newsletter writers get first drafts of curated roundups.

The Limitations of Manual Reading

Reading and summarizing ten long articles can take an afternoon. Automated summaries give you the gist in minutes, so you only read in full what really matters.

Architecture Overview

Manual Trigger → HTTP Request (index) → HTML extract links → Split Out → Limit → HTTP Request (article) → HTML extract text → Summarization Chain (OpenAI + text splitter) → Merge → Set. For larger content pipelines, see our AI news automation use case or explore Vyrade's Automation Blueprint.

Conclusion

This workflow is the simplest way to see scraping and AI working together in n8n. Point it at any index page, adjust the selector and prompt, and you have a reusable research assistant. Always respect each site's terms and robots.txt; Google's robots.txt introduction is a good primer.

Frequently Asked Questions

It works on most static pages. You may need to change the CSS selectors for each site.