Quick Overview
This workflow collects a landing page URL via an n8n Form, scrapes and extracts the page text, uses an OpenAI-compatible chat model (deepseek-v4-flash) to generate a CRO score, roast, and fix list, logs results to Google Sheets, emails a report with Gmail, and sends low-score alerts to Telegram plus a weekly email digest.
How it works
- Receives a landing page URL and recipient email through an n8n Form submission.
- Normalizes the inputs, validates the URL and email, and sends a Gmail rejection email if the submission is invalid.
- Fetches the landing page HTML via HTTP, strips it down to clean text, and limits the extracted content size for analysis.
- Sends the cleaned page content to an OpenAI-compatible chat model (deepseek-v4-flash) to return a JSON audit with an overall score, sub-scores, a roast summary, and 10 prioritized recommendations.
- Parses the AI JSON, derives a letter grade, formats a plain-text report, and prepares a structured audit log record.
- Appends the audit record to a Google Sheets “Audits” tab, emails the full report to the submitter via Gmail, and posts a Telegram alert when the score is below 75 before showing a completion message.
- Every Monday at 9:00, reads the audit log from Google Sheets, computes last-7-days stats, uses the same OpenAI-compatible model to write a short digest, and emails it via Gmail.
Setup
- Add an OpenAI API credential for the two AI steps and ensure the selected model (deepseek-v4-flash) is available in your provider/account.
- Connect Google Sheets OAuth2 credentials and set the spreadsheet document ID and “Audits” sheet name where audit rows are appended and later read for the weekly digest.
- Connect Gmail OAuth2 credentials and update the recipient addresses used for the audit report, invalid-input notice fallback, and weekly digest email.
- (Optional) Connect Telegram credentials and update the chat ID used for low-score alerts.
- Publish the n8n Form and share its URL so users can submit landing page audits.