n8n Workflow: Scrape Portfolio Credits with OpenAI & Google Sheets

Scrape portfolio case-study pages from your Google Sheets and extract project credits with OpenAI

This n8n workflow reads a list of portfolio URLs from Google Sheets, fetches each page, converts the HTML to clean text, and uses an OpenAI chat model to extract structured credits (Agency, Brand, Project Name). It then writes both results and errors back to Google Sheets—marking each input row as Done so you can run it repeatedly.

What this workflow does

• Run manually to start the process.
• Load input rows from your Google Sheets Input tab and filter out URLs already marked as Done.
• For each remaining row, it sets the Company and Page URL fields, then fetches the portfolio page HTML using browser-like headers.
• Validate responses and log a fetch error to Google Sheets when retrieval fails.
• If the page loads, strip HTML to plain text (with truncation) to prepare content for analysis.
• Send the cleaned page text to an OpenAI chat model extraction agent to return a structured list of projects and credits.
• Write outputs to the Google Sheets Output tab:
  • Extracted project credits when extraction succeeds
  • An AI extraction error entry when extraction fails
• Update the input row status to Done and wait briefly before moving to the next URL.

Use cases

• Automatically compile agency/brand/project credit data from case-study pages into a structured sheet.
• Batch-audit partner portfolios by running one job and reviewing Errors for failed pages.
• Maintain a clean, repeatable pipeline for SaaS teams tracking credits across many websites.

Technical details

• Google Sheets (Input/Output tabs): reads URLs, appends extracted projects, and writes fetch/AI errors.
• OpenAI chat model: extracts structured credits from cleaned page text.
• n8n nodes & controls: if, set, code, wait, filter, sticky note.

Best used when your spreadsheet schema includes: Company, Website, Page URL, Status (Input) and Agency, Brand, Project Name, Full Page Content, Error (Output).