Turn chat questions into clean web page summaries you can use right away. An AI agent reads your request, fetches the page, strips noise, and returns short Markdown that is easy to scan and share. Great for marketing and research teams that need quick page reviews for briefs and competitor checks.

It starts from a manual chat message or as a called subworkflow. The agent uses OpenAI to decide when to call a web fetch tool with a query like ?url=site&method=full or simplified. The flow downloads the page, keeps only the body, removes scripts, ads, media frames, and comments, then turns the rest into Markdown. A size limit protects token use, and errors come back with clear tips so the agent can try again.

You only need an OpenAI API key and n8n credentials. Set the default max page size and choose simplified mode when you do not need links. Expect faster research, lower token costs, and clean outputs for content briefs, SEO notes, and competitor snapshots. You can also call this from other workflows to standardize web capture across your team.

• Manual chat trigger starts an AI guided web fetch from your prompt
• ReAct agent with OpenAI chooses when and how to call the web tool
• Query string parser turns ?url and method inputs into structured settings
• HTTP request encodes the URL, tolerates site errors, and keeps the run alive
• HTML body extraction and tag cleanup removes scripts, iframes, media, and comments
• Simplified mode replaces href and src with placeholders to shrink content
• Markdown conversion keeps headings and lists while removing heavy HTML
• Length check returns a clear error when a page is too long
• Callable by other workflows through Execute Workflow Trigger for reuse

• Reduce manual work from 30 minutes to 2 minutes per page
• Streamline page research by about 60% with one chat request
• Cut token usage by up to 20% with simplified mode that removes links and image URLs
• Improve clarity by removing scripts, ads, and media frames for cleaner analysis
• Prevent runaway costs with a strict max page size control
• Connect AI decisions and web content in one chat driven flow

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with OpenAI. See the Tools Required section above for links to create accounts with these services.
3. In the n8n credentials manager, create a new OpenAI API Key credential. Visit the OpenAI API Keys page, create a key, paste it into n8n, and save.
4. Open the OpenAI Chat Model node and select your OpenAI credential. Keep the default settings or adjust temperature and penalties as needed.
5. Open the ReAct AI Agent node and confirm the OpenAI Chat Model is linked and the HTTP_Request_Tool is listed as a tool.
6. Review the HTTP_Request_Tool node description so you know the correct input format: ?url=VALIDURL&method=full or simplified. No credentials are needed for this tool.
7. Open the CONFIG Set node and adjust query.maxlimit to set the maximum allowed page size for output.
8. Start a test chat. Ask for a page review or summary. The agent will decide the method and call the tool. You can also provide a direct tool input like ?url=https://example.com&method=simplified.
9. Check the Send Page Content node output. You should see clean Markdown in the page_content field. If you see INVALID action_input, change the tool input to a proper query string.
10. If you get ERROR: PAGE CONTENT TOO LONG, lower the maxlimit, switch to simplified mode, or choose a smaller page.
11. To reuse this in other automations, add an Execute Workflow node in another workflow, select this workflow, and pass a query string so it returns cleaned Markdown.