Instead of just making pretty pictures, I used it to turn *any* YouTube video into a clean, data packed infographic that posts straight to LinkedIn. Here is how the system works step by step: 1. Capture the video and context - Submit a YouTube link through an n8n form - Add special instructions like what to highlight or which sections to focus on 2. Extract the transcript - Use Apify’s YouTube Transcript scraper - Pull the full transcript of the video in one go 3. Turn the transcript into an infographic prompt - Clean the raw transcript with a small code node - Use OpenAI to convert it into a structured infographic prompt - Define sections like title, key points, risks, players, and layout 4. Generate the infographic with Nano Banana Pro - Call Nano Banana Pro through Kie.ai using HTTP requests - Feed the optimized prompt into the model - Get a professional infographic image that looks like a designer built it 5. Write the LinkedIn caption with AI - Use another OpenAI step to write a hook and caption - Format it for LinkedIn so it is easy to read and saves people time 6. Auto post to LinkedIn - Download the image from Kie.ai - Attach it to a LinkedIn post straight from n8n - Use the AI caption as the post text The result You paste a YouTube link. The automation gives you a ready to post infographic plus a high converting caption. I linked the full tutorial and blueprint in the video so you can copy the system and tweak it to your own content. Subscribe if you want more practical AI and automation builds like this. https://youtu.be/kjbvMOudsvM?si=8018imAjN2Vah6EH

If you are building AI agents in n8n, there are a few features that can take them from basic to genuinely useful. A lot of people only use the default setup, and they miss out on things that make agents more reliable, more accurate and much more powerful. In this video I break down five features that every agent should be using. Some of them are simple. Others are overlooked because most tutorials skip them. I walk through everything using a real example, which is my budget tracking agent that logs expenses into a spreadsheet through Telegram. Here is what I cover: 1. How to use a system prompt the right way 2. How tool calling works and how to let the agent fill parameters on its own 3. How to use structured output so your agent returns clean and predictable data 4. How simple memory works and when to store chat history 5. How to delegate tasks to a sub agent and avoid long messy prompts These features apply to support agents, CRM update bots, appointment setters and any workflow that relies on AI reasoning. If you want more agent breakdowns and practical builds, make sure you subscribe.

Google just released Gemini 3 and with it their new image model Nano Banana Pro. The results are crazy realistic and the price is surprisingly low. So I built a full image generation system in n8n that lets you create images through a simple chat interface. The bot can generate three types of images. You can create an image from text. You can use an existing image as a reference. Or you can upload multiple images and combine them into a brand new one. I walk through how the whole system works: 1. How to connect Nano Banana Pro through Kie.ai using HTTP requests 2. How to build a clean n8n form that collects prompts and file uploads 3. How to enhance prompts with AI before sending them to the model 4. How to use Google Drive as a free image host for the uploaded files 5. How to handle the async generation loop until the image is done 6. How to send the final image back through a Telegram chat The bot can make infographics, edits, memes or realistic image composites. And you can run the whole thing from your phone. If you want more builds like this, make sure you subscribe. https://youtu.be/sJsHQ3J1erI?si=8KcVFwNno7R3gFk0

I’m looking for honest feedback on a platform idea I’m exploring. I suspect this group may have some educated feedback for me. The Problem: "Doom Scrolling Guilt" - I struggle with this. We all doom scroll on TikTok/Reels. It feels good in the moment, but we often feel like trash afterward because we didn't learn anything. On the flip side, we want to learn, but 2-hour video courses or 40-page eBooks feel like "homework." The Solution: I want to build a "TikTok for Carousels" (I know, "TikTok for X" is a cliché, but it describes the UX perfectly). Think of the educational PDF carousels you see on LinkedIn (5-10 slides, bite-sized tactical value), but in a dedicated mobile app designed for binge-learning, not networking. - UX: Vertical scroll to discover new topics → Horizontal scroll to read the specific lesson. - The Content: Starting with Sales & Marketing (high-value, tactical content). - For Creators: Currently, carousel creators get likes but $0. This platform lets them gate content. Proposed Monetization:"Freemium Slides." - Slides 1-3 are free (The Hook). - Slide 4-10 are locked. - User pays to unlock the rest + get downloadable templates. I need your help with 3 specific debates: 1. The Psychology: Am I giving people too much credit? Do people actually want to learn in their spare time, or do they just say they do but actually prefer mindlessly watching cat videos? 2. Ads vs. Micropayments: As a user, would you rather: 3. A) Pay small amounts (e.g., $0.50) to unlock a specific guide?B) Pay a flat sub (e.g., $5/mo) for unlimited access?C) Get it for free, but have Ads inserted between slides? 4. The Format: Is the "Tiktok Carousel" format actually a good learning method, or is there another delivery format you would recommend? Roast the idea - I’d rather know it’s bad now than in 6 months. Will be built with Supabase, n8n, and NextJS. Thanks!

A lot of AI conversations are still stuck on “How smart is the model?”In real businesses, the better question is: “How dumb‑proof is the system around it?”AI can already write code, analyze deals, and draft emails at a level that would’ve seemed wild a few years ago. But in production, most of the wins (or disasters) don’t come from the model – they come from the guardrails, data, and decisions humans wrap around it.The pattern that keeps showing up:Smart data in, smart behavior out. Garbage in, and the AI will confidently amplify your worst assumptions.Clear guardrails = fewer “oops” moments with customers, cash, or compliance.Human‑in‑the‑loop at key checkpoints turns AI from a black box into a power tool you can actually trust.That’s why the interesting frontier (at least for me) isn’t “let AI run everything.” It’s: “Where can AI do 80–90% of the heavy lifting, and where do we design the system so it’s almost impossible to break something important?”In my world, that looks like using AI to reason about complex financial and tax logic in commercial real estate, while being obsessed with:What data it’s allowed to seeWhat decisions it’s allowed to touchWhere a human must review anything that affects real money or long‑term trustYou don’t need to reveal your prompts, your data sources, or your secret sauce to have this conversation. You can still share the principle:“AI is as smart as we make it dumb‑proof with data, guardrails, and humans in the loop.”Curious how others here are handling this: Where do you let AI run on rails, and where do you insist on a human checkpoint before anything goes live?.