Templates / Create weekly AI research briefings with Decodo, OpenAI, and Gmail

Create weekly AI research briefings with Decodo, OpenAI, and Gmail

Unknown

by Abdullah Alshiekh ✓ · 50 views · created 2025-12-23 · Market Research, AI Summarization

Unknown. This template uses community nodes we can't verify statically. Uses community nodes that can't be verified statically (unverified community packages are off by default in 3.0). Static check, last run 2026-10-19.

Community nodes

These come from third-party npm packages. In n8n 3.0 unverified community packages are off by default (N8N_UNVERIFIED_PACKAGES_ENABLED=false).

Nodes used (11)

Summarization ChainAI AgentGmailOpenAI Chat ModelTelegram

What it does

What Problem Does It Solve? - Staying up-to-date with AI and LLM developments requires reading dozens of articles every week. - Manual research is time-consuming and often leads to “information overload” or reading low-quality clickbait. - Important technical breakthroughs often get buried under marketing fluff. This workflow solves these by: - Leveraging Decodo to instantly find and scrape high-quality organic articles (automatically filtering out YouTube/video noise). - Using AI to read and summarize every article individually. - Using an "Analyst Agent" to score news by relevance and write a single, high-quality intelligence report. How to Configure It Decodo & API Setup - Decodo: Connect your Decodo credentials. This is the core engine that handles the high-precision Google Search and content scraping. - OpenAI: Connect your OpenAI API key (GPT-4o or 4.1-mini recommended for best analysis). - Gmail: Connect a Google Service Account or Gmail OAuth to send the emails. Search Configuration - Open the Set Search Config node. - Edit the searchquery value to match your niche (e.g., "Latest Large Language Model benchmarks" or "Generative AI in Healthcare"). How It Works - Trigger: The workflow wakes up once a week (customizable). - Search (Powered by Decodo): It searches Google using Decodo's organic results filter to ensure only high-quality reading material is selected. - Scraping: It visits every URL found and extracts the raw text, cleaning up HTML tags. - Summarization: An LLM reads each article individually to extract key technical points. - Analyst Agent: Reviews all summaries, assigns a "Relevance Score", and compiles the final newsletter. - Delivery: The final report is emailed to you immediately. Customization Ideas - Change the topic to any industry (Crypto, Finance, Sports). - Swap the AI model for Claude or DeepSeek. - Log the summaries into a Notion database. If you need any help Get In Touch

View the original on n8n.io → Check your own copy

Get the weekly “what broke” email

New templates, newly broken ones, and n8n release changes. One short email a week.

Form not loading? Subscribe here.