Summarize web pages from Google Sheets using OpenAI GPT-4o-mini

Quick overview

This workflow reads a list of web page URLs from Google Sheets, fetches each page, and uses OpenAI Chat Completions to generate a factual title, short summary, key points, topic, and confidence score, then writes the results back to the same sheet.

How it works

1. Starts when you manually execute the workflow.
2. Loads run settings (maximum pages per run and the OpenAI model) and reads rows from the pages tab in Google Sheets.
3. Selects up to the configured limit of rows that have a URL but no existing summary, normalizes/validates the URL, and skips duplicates.
4. Fetches each web page over HTTP and extracts plain text (including the HTML title) for summarization.
5. Sends the page text to the OpenAI Chat Completions API with instructions to return strict JSON containing title, summary, key points, topic, and confidence.
6. Parses the OpenAI response, marks rows as done when a real summary is produced (or failed with blank fields if the page or API call fails), and updates the matching rows in Google Sheets by row_number.

Setup

1. Create or use a Google Sheet with a tab named pages that includes columns row_number and url, plus title, summary, key_points, topic, confidence, status, and summarised_at for the workflow to fill.
2. Add Google Sheets OAuth2 credentials and paste your Google Sheet URL into both Google Sheets nodes (read and update).
3. Add an OpenAI API credential for the OpenAI HTTP request node.
4. Optionally adjust max_per_run (1–100) and choose an allowed model (for example gpt-4o-mini) in the configuration step before running.