Automate AI News: Gemini Classify RSS to LinkedIn Carousels (n8n, PostNitro)

Automate AI news curation + LinkedIn carousel drafts—powered by n8n, Gemini, and PostNitro

This n8n workflow automatically polls TechCrunch for new AI-related articles, uses Google Gemini to classify and analyze them, then sends the results to PostNitro to generate a scheduled LinkedIn document carousel draft—complete with Gmail success/error notifications.

What this workflow does

• Polls the TechCrunch RSS feed every minute and reads newly updated items.
• Limits processing to the first 10 articles from each feed update.
• Uses Google Gemini (PaLM) to classify each article as AI-relevant or not, assign a category, and score importance using structured output.
• Filters out non–AI-relevant items automatically, so only relevant stories proceed.
• Runs a second Gemini step to produce a structured analysis for each relevant article, including impacts and a recommended carousel angle.
• Sends the article URL to PostNitro to generate a carousel design draft.
• Schedules the post on LinkedIn as a LinkedIn document for the next day.
• Sends a Gmail email notification confirming the scheduled draft, or an error email if PostNitro generation fails.

Use cases

• SaaS marketing teams who want consistent LinkedIn carousels from AI news without manual curation.
• Automation engineers building an end-to-end RSS → LLM classification → carousel publishing pipeline.
• Founders/operator dashboards who need fast, structured AI news selection and scheduling.

Technical details

• n8n nodes/blocks: RSS Feed Read Trigger, Limit, If, n8nn8n-nodes-langchainchain (LLM integration), PostNitro API, Gmail, and supporting nodes like Sticky Note.
• Requires configuring Google Gemini (PaLM) credentials for both the classifier and analyst steps.
• Requires configuring PostNitro API credentials (brand ID, preset ID, template ID, and selected account ID) for carousel creation and scheduling.
• Requires Gmail OAuth2 plus recipient email and success/failure message content.