Generate SEO-optimized blog content with Google Gemini, RSS feeds and Telegram

Workflow Objective

This n8n workflow automates the entire content creation process by monitoring specified RSS feeds for new articles. It then leverages Google Gemini AI to generate comprehensive, SEO-optimized blog posts inspired by these articles, creates unique cover images, and distributes the final content as a JSON file to stakeholders via Telegram. The primary goal is to automate the end-to-end content pipeline, saving significant time and ensuring a consistent output of high-quality content.

Step-by-Step Breakdown

1. Monitor News Sources (RSS Triggers): The workflow is triggered periodically (e.g., hourly, weekly) by multiple RSS Feed nodes that monitor sources like "Search Engine Journal" and "Tech Crunch" for new publications.

2. Prevent Duplicate Content (Deduplication): For each new article fetched from the RSS feeds, the workflow checks an AWS DynamoDB database to see if the article's URL has been processed before. If the link already exists in the database, the process for that item is halted, and a debug notification is sent to Telegram via the "Telegram Debugger" node. This prevents the generation of duplicate content.

3. AI-Powered Content Generation (Gemini Content Generation): If the article is new, its link is passed to a Google Gemini node. Using a highly detailed and structured prompt, Gemini generates a complete blog post in a specific JSON format. This output includes a title, meta description, SEO-friendly slug, a descriptive prompt for generating a cover image, and the full markdown body of the article (including an introduction, subheadings, conclusion, FAQ section, etc.).

4. Data Cleaning and Parsing (JSON Parser): The raw text response from the AI is processed by a "Code" node. This custom script cleans the output—removing markdown code blocks, fixing potential syntax errors—and reliably parses it into a valid JSON object, ensuring the data is clean for subsequent steps.

5. Image Generation and Cloud Storage: The imagegenerationprompt from the parsed JSON is sent to another Google Gemini node configured for image generation, creating a 1200x630 cover image for the blog post. The newly created image is renamed using the slug. Finally, the image is uploaded to a cloud storage service like Cloudflare R2. If the upload fails, an error message is sent to Telegram.