I built a custom integration that lets n8n's AI Workflow Builder use Claude models (Sonnet 4.5, Opus 4.5) without paying per-token API costs. It's working. And it changes everything about how I build automations now. --- 🎬 Watch the Demo (Video Above) In this video, you'll see me: - Configure the connection in n8n's settings - Select my AI model - Generate workflows with Claude — completely free No API key billing. No usage anxiety. Just pure building. --- 📖 Why I Built This When you're designing AI workflows, the real work is iteration: - Test a prompt → tweak it → test again - Add a node → see if the AI gets it - Refine → regenerate → compare results With pay-per-token billing, there's always that voice in your head: "Is this test worth it?" That friction kills creativity. I wanted to remove it completely. --- ✅ What It Looks Like Now I added a new section in n8n's Personal Settings: "AI Builder Connection" - One-click configuration - Choose your model: - Claude Sonnet 4.5 - Claude Sonnet 4.5 (Thinking) - Claude Opus 4.5 (Thinking) - Gemini 3 Pro - Connect once, build unlimited The AI Workflow Builder now feels like a creative partner, not a metered service. --- 🧠 What This Means for Building With this setup, I can: - Generate 50+ workflow variations in one session - Experiment with Opus-level reasoning for complex automations - Iterate freely without watching a billing dashboard - Actually enjoy the building process again It's the difference between "let me think if this is worth testing" and "let's just try it and see." --- 💬 Question for You I'm curious — how do you currently handle AI costs when building workflows? - Do you limit your experiments? - Use cheaper models even when you need more power? - Just accept the cost as part of the process? Drop your thoughts below 👇 --- 🚀 What Are You Building? I always love hearing what others are automating with AI: - Lead gen pipelines? - Content workflows? - Internal tools? - AI agents?

Hey everyone 👋 I’m doing a bit of research around n8n usage and real-world pain points, and I’d love your input. If you could add one new n8n node (or significantly improve an existing one), what would you choose and why? Is there something you constantly have to hack around with HTTP requests, custom code, or external tools? No right or wrong answers here — I’m genuinely trying to understand where the biggest gaps are for people building automations daily. Looking forward to your thoughts 👀

Over the weekend I made a Bash script-based workflow to automatically check if there is a new N8N release available from the official N8N repository. Details as follows: - Executed using `cron` - Executes at 3am NZDT (Local time) - Delivers a desktop notification using Gnome - Sends an email to the nominated address - Logs the action, Success/Fail to a separate log - Tracks the release version in a separate file I had built it out to a complete system but that was essentially overkill, so I stripped it back: - quickstart.txt - setup.sh - install-cron-v2.sh - n8n-release-monitor-v2.sh - README-distribution.md TODO #1: Convert this into an N8N Workflow this week. TODO #2: Create a `rm -f *` for this workflow Note #1: This was built for and tested on an Ubuntu Linux 25.04 Desktop & Server. Note #2: Check the scripts before running and caveat emptor rules apply.

A have Create a custom n8n node for working with PDF documents using the pdf-lib library. This node allows you to fill PDF forms, create new PDFs, and merge multiple PDFs directly in your n8n workflows. 🎯 Features Fill PDF Forms - Add text overlays to existing PDF documents with full positioning control Fill Form Fields - Fill actual interactive PDF form fields (text, checkboxes, radio buttons) Extract Form Fields - Discover all form fields, types, values, and properties from PDFs ✨ NEW in v0.4.0 Add Images - Embed PNG and JPEG images into PDFs with positioning and rotation Create PDFs - Generate new PDF documents from scratch Merge PDFs - Combine multiple PDF files into one with support for multiple sources JavaScript Code Mode - Programmatically define fields using JavaScript code French Character Support - Full UTF-8 support for international characters Flexible Input - Works with URLs or binary data Customizable - Configure fonts, colors, positions, sizes, and rotations 🚀 Usage Fill PDF Form Perfect for filling tax receipts, invoices, or any form-based PDFs: Select Fill PDF Form operation Choose PDF source (URL or Binary Data) Add text fields with coordinates and styling Output as binary data Example Use Case: French Tax Receipt (Reçu Fiscal) Load template PDF from URL Add donor name, address, amount, date Generate personalized receipt Create PDF Create new PDF documents from scratch: Select Create PDF operation Choose page size (A4, Letter, or Custom) Add text fields with positions and styling Output as binary data Extract Form Fields ✨ NEW Discover all form fields in a PDF before filling: Select Extract Form Fields operation Choose PDF source (URL or Binary Data) Get JSON output with all field information Returns: Field names and types (text, checkbox, radio, dropdown, list, button, signature) Current values and properties Max lengths, options, read-only status PDF metadata (page count, title, author) Perfect for:

This agent listens to selected sources of data: 1. Twitter accounts 2. Newsletters 3. Websites Every day it aggregates all that data and saves important info to the Google Sheet I built this automation for myself to track latest AI trends and tool Giving it away to you for free, enjoy! n8n templates is attached Google Sheets template: https://docs.google.com/spreadsheets/d/1K10bvT-mUt1OjCNQSjRgUCuI0BgZcvdpRFpdRLIfUS8/edit?gid=1069367716#gid=1069367716