Build Personalized AI Newsletters with QWEN & Gemma Using RSS Feeds and Gmail

How it works

• Collects articles from your preferred RSS feeds.
• Rates and tags each article using an AI model (e.g., QWEN 14B-s4), filtering for relevance and quality.
• Summarizes high-rated articles with a language model (e.g., Gemma3 4B) for quick, digestible reading.
• Checks for duplicates to avoid sending the same article twice.
• Formats and sends the top articles as an HTML newsletter via Gmail, using OAuth2 authentication.
• Stores records in a Postgres database, tracking which articles have been sent and their ratings.

Requirements

• Postgres Account
• AI Models (if you work localy use Ollama)
• In the cloud you have to change Ollama node to your prefered Model Node
• RSS Feed of your desire
• Google Auth2, if you want to use Gmail

Recommendations

• Use n8n local version for this workflow
• Here are some more informations: https://github.com/falks-ai-workbench/n8n_newsletter