Leon van Zylπ€ Ollama just launched their own ChatGPT-style interface - no more terminal commands needed to run local AI models securely on your machine!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Set up Ollama's brand new chat interface in minutes πΉ Download and run AI models directly from the UI (no terminal required) πΉ Upload documents for AI analysis and Q&A sessions πΉ Test cutting-edge reasoning models like DeepSeek R1 πΉ Create custom AI characters with personalized system prompts πΉ Optimize context length for better document processing
π― PERFECT FOR:
β Privacy-conscious users wanting local AI control β Beginners intimidated by command-line interfaces β Developers testing different AI models locally β Open Web UI users looking for simpler alternatives β Anyone wanting free, offline AI capabilities
00:00 - Ollama's new chat app introduction 01:04 - Download and install Ollama setup 01:32 - Model downloading through UI interface 02:29 - Chat functionality and document uploads 03:22 - Context length settings optimization 04:16 - DeepSeek R1 reasoning model demo 05:00 - Vision capabilities with multimodal models 05:43 - Creating custom models with system prompts 07:15 - Testing custom Mario character model 08:11 - Feature requests and final thoughts
Finally eliminates the complexity of running local AI models - now anyone can have ChatGPT-style conversations with open-source models running completely offline and privately on their own machine.
Ollama Just Released Their Own App (Complete Tutorial)Leon van Zyl2025-07-31 | π€ Ollama just launched their own ChatGPT-style interface - no more terminal commands needed to run local AI models securely on your machine!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Set up Ollama's brand new chat interface in minutes πΉ Download and run AI models directly from the UI (no terminal required) πΉ Upload documents for AI analysis and Q&A sessions πΉ Test cutting-edge reasoning models like DeepSeek R1 πΉ Create custom AI characters with personalized system prompts πΉ Optimize context length for better document processing
π― PERFECT FOR:
β Privacy-conscious users wanting local AI control β Beginners intimidated by command-line interfaces β Developers testing different AI models locally β Open Web UI users looking for simpler alternatives β Anyone wanting free, offline AI capabilities
00:00 - Ollama's new chat app introduction 01:04 - Download and install Ollama setup 01:32 - Model downloading through UI interface 02:29 - Chat functionality and document uploads 03:22 - Context length settings optimization 04:16 - DeepSeek R1 reasoning model demo 05:00 - Vision capabilities with multimodal models 05:43 - Creating custom models with system prompts 07:15 - Testing custom Mario character model 08:11 - Feature requests and final thoughts
Finally eliminates the complexity of running local AI models - now anyone can have ChatGPT-style conversations with open-source models running completely offline and privately on their own machine.
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#ollamaClaude Sonnet 4.5 + Claude Code 2.0: Agentic Coding MasterclassLeon van Zyl2025-10-03 | Learn how to master Claude Code 2.0 and Claude Sonnet 4.5 - Anthropic's latest coding model that's revolutionizing AI-powered development. This comprehensive crash course covers everything from basic setup to advanced features like subagents, MCP servers, custom commands, background tasks, checkpointing, and GitHub integration. Whether you're using the CLI tool or VS Code extension, you'll discover how to leverage Claude Code for maximum productivity in your coding projects.
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
β° TIMESTAMPS: 00:00 - Claude Sonnet 4.5 release and new features overview 01:03 - Setting up VS Code for Claude Code 01:51 - Installing the Claude Code VS Code extension 02:18 - Login and authentication setup 03:08 - Installing Claude Code CLI tool 04:33 - Windows vs WSL installation differences 05:09 - Basic Claude Code usage and modes 06:16 - Bypass permissions and thinking modes 07:48 - Adding images to prompts 09:00 - Managing conversation history 09:36 - Understanding memories and project rules 10:23 - Making changes to existing projects 11:59 - Creating custom commands 14:05 - Checkpointing and rewind feature 16:20 - Background tasks and bash management 18:10 - Creating and using subagents 20:38 - Setting up MCP servers with Shad CN 25:02 - Adding MCP servers via CLI 26:34 - Hooks for notifications and automation 29:46 - GitHub integration for remote development 32:43 - Series wrap-up
#claudecode #claudesonnet45 #anthropicSpec Kit: How to Build Production-Ready Apps with AI AgentsLeon van Zyl2025-10-01 | Learn how to build production-ready applications with coding agents using Spec Kit - a structured workflow from GitHub that brings spec-driven development to AI coding assistants like Claude Code, Cursor, Windsurf, and Codex. This tutorial walks through the complete Spec kit workflow including Constitution setup, feature specification, planning, task breakdown, and implementation using a Next.js expense tracking app demo. You'll discover how to use proper Git branching, test-driven development, and structured prompting to get consistent, high-quality results from your coding agent instead of broken features and bugs.
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
β° TIMESTAMPS: 00:00 - Why coding agents need structure 00:46 - What is Speckit and spec-driven development 02:26 - Core workflow components overview 04:36 - Creating a new Next.js project 05:07 - Installing Speckit with UV 06:59 - Constitution: Setting project principles and standards 09:09 - Using Speckit across different coding agents 12:00 - Specify: Defining your feature requirements 14:00 - Clarify: Answering agent questions to refine specs 16:41 - Plan: Creating technical implementation plan 19:20 - Tasks: Breaking plan into actionable steps 20:42 - Implement: Building the expense tracker app 22:30 - Test-driven development workflow in action 24:58 - Manual testing and reviewing the completed app 26:00 - Git workflow: Creating pull requests and merging 26:50 - Adding new features with the same workflow
#speckit #claudecode #codexI Built the Ultimate AI Podcast System in n8n (No Code Required)Leon van Zyl2025-09-29 | Learn how to build an AI podcast generator using n8n workflow automation and ElevenLabs V3 text-to-dialogue technology. In this comprehensive tutorial, you'll discover how to create natural sounding multi-speaker conversations by scraping web content with FireCrawl, generating dialogue with OpenAI GPT, and producing realistic AI voices using ElevenLabs V3 model. We'll cover the complete workflow setup, production deployment with Hostinger VPS, webhook integration, and building a custom web application with Lovable.
π Save on n8n Hosting with Hostinger: hostinger.com/leon (Use code: LEON for 10% off)
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
β° TIMESTAMPS: 00:00 - Introduction: Building AI podcast generator with N8n and ElevenLabs 00:27 - Demo: Natural multi-speaker conversation generation 01:27 - Setting up N8n workflow and manual trigger 02:16 - Configuring FireCrawl for web content scraping 03:00 - Creating FireCrawl API key and connection setup 04:00 - Building conversation generator with OpenAI GPT 05:00 - Understanding ElevenLabs V3 structured output format 06:00 - ElevenLabs service overview and V3 model capabilities 07:16 - Selecting and configuring voice IDs for speakers 08:00 - Creating system prompt with personality definitions 09:00 - Testing conversation generation workflow 09:17 - Setting up ElevenLabs HTTP node for audio generation 10:00 - Creating ElevenLabs API credentials and authentication 11:00 - Testing initial audio output and identifying improvements 12:00 - Implementing ElevenLabs V3 prompting techniques 13:00 - Enhanced audio results with natural speech patterns 13:24 - Deploying N8n workflow to Hostinger VPS production 15:00 - Setting up production environment and credentials 16:00 - Configuring N8n cloud instance and license activation 17:00 - Production workflow setup and testing 18:00 - Creating webhook triggers for external integration 19:00 - Testing webhook with Postman API calls 20:00 - Handling multiple trigger types and URL parameters 21:00 - Activating production workflow for external access 21:50 - Building custom web app integration with Lovable 22:50 - Conclusion and next steps
#n8n #elevenlabs #aipodcastRun n8n with Docker Model Runner Locally (Free AI Models)Leon van Zyl2025-09-23 | Learn how to use Docker model runner with N8N instead of Ollama for running free open source AI models locally. This tutorial shows you how to set up Docker Desktop, download models like GPT OSS, and integrate them with your N8N workflows using GPU acceleration. You'll discover how to configure the OpenAI-compatible API, set up embedding models for vector databases, and create a complete AI agent setup with Postgres database integration for knowledge base queries.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
β° TIMESTAMPS:
0:00 - Series recap: N8N with free models and Ollama setup 0:29 - Why use Docker model runner instead of Ollama 0:43 - Docker model runner benefits: GPU acceleration and easy integration 1:08 - Installing Docker Desktop and downloading models 1:37 - Accessing models through Docker Desktop interface 1:54 - Command line access to Docker models 2:20 - OpenAI-compatible API for external applications 2:52 - Enabling Docker model runner in settings 3:28 - Integrating Docker model runner with N8N 3:58 - URL configuration for Docker vs local N8N setup 4:38 - Testing the N8N integration with Docker models 5:03 - Setting up embedding models for knowledge base 5:42 - Testing vector database queries with Anthropic invoice example
#n8n #docker #aiHow to Run n8n Locally with Free AI & Memory (No Code)Leon van Zyl2025-09-22 | In this n8n tutorial you will learn how to use n8n for FREE.
Learn how to set up n8n completely for free using Docker Desktop, Ollama AI models, and PostgreSQL database. This comprehensive tutorial shows you how to create powerful AI chatbots and agents without any subscription costs. You'll discover how to use free open-source AI models through Ollama, set up persistent conversation memory with PostgreSQL, build custom knowledge bases from Google Drive documents, and expose your n8n instance publicly using NGROK for WhatsApp and Telegram integration.
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
β° TIMESTAMPS: 00:00 - Introduction: Free N8n setup overview 00:27 - Installing Docker Desktop and WSL setup 01:42 - Downloading and running N8n container 02:27 - Creating your first N8n workflow and AI agent 04:16 - Setting up Ollama for free AI models 06:02 - Connecting Ollama to N8n chat model 07:08 - Adding conversational memory to your agent 08:38 - Setting up PostgreSQL for persistent storage 12:04 - Building knowledge base with Google Drive integration 13:00 - Creating Google Cloud project and OAuth setup 16:00 - Document processing and vector embeddings 19:38 - Making N8n publicly accessible with NGROK 21:40 - Testing public chat interface and wrap-up
#n8n #ollama #dockerHow to Add Human-Sounding AI Dialog (ElevenLabs V3)Leon van Zyl2025-09-20 | In this video you will lean how to use Eleven v3 - a powerful new model from ElevenLabs - to add natural sounding AI voices to your applications. In this example, we create a podcast style dialog between multiple speakers using ElevenLabs v3. You will learn how to add the Elevenlabs SDK to your Nextjs project, generate the ElevenLabs API key and use the the Text to Dialog method.
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
β° TIMESTAMPS: 00:00 - Series recap: From article scraping to AI conversations 00:42 - 11 Labs V3 demo: Natural multi-speaker capabilities 01:15 - Installing 11 Labs SDK and setting up project 01:50 - Getting your 11 Labs API key (step-by-step) 02:30 - Environment setup and SDK installation 03:00 - Using coding agents to implement the solution 04:05 - Exploring voice options: Best voices for V3 05:00 - Setting up voice IDs for excited and skeptical hosts 05:50 - First audio generation test and results 07:00 - Improving dialogue with 11 Labs prompting guide 08:00 - Final testing: Natural conversation with interruptions 09:20 - Series wrap-up and future project ideas
#elevenlabs #codex #gpt5Donβt Ship Another AI App Until You Try This Streaming TrickLeon van Zyl2025-09-19 | Learn how to stream structured output in real-time using Vercel's AI SDK instead of waiting for complete LLM responses. This step-by-step tutorial shows you how to build apps with immediate user feedback as data generates.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvzAdd Web Scraping to Any App in Just 5 Minutes with FirecrawlLeon van Zyl2025-09-18 | Learn how to integrate powerful web scraping into any application using the Firecrawl SDK in this comprehensive step-by-step tutorial. This video covers the complete implementation process from documentation research to production-ready code, including secure API key management with environment variables, real-time content extraction, and proper error handling. Perfect for developers building AI applications, data collection tools, or any project requiring reliable web scraping capabilities. We demonstrate scraping live websites, processing markdown content, and integrating the scraped data into your app's user interface. Whether you're building chatbots, content analyzers, or automated research tools, this tutorial provides the foundation for robust web scraping that works across any website without getting blocked. Includes best practices for project documentation, dependency management, and maintaining clean, scalable code architecture that follows industry standards.
β° DETAILED TIMESTAMPS:
00:00 - Previous video recap: AI podcast generator UI 00:30 - Why documentation matters for every project 01:42 - Finding and using Firecrawl documentation 02:11 - Saving technical docs locally vs MCP services 03:05 - Implementing web scraping feature requirements 04:23 - Creating Firecrawl API key and account setup 05:31 - Environment variable security best practices 06:28 - Testing web scraping with real OpenAI article 07:07 - Code review: API routes and implementation details 08:02 - Creating git checkpoint for feature completion
Learn professional development practices that coding agents struggle with - proper planning, architecture design, and model flexibility that scales to enterprise applications.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#claudecode #codex #vibecodingBuild AMAZING UIs with AI Agents in MinutesLeon van Zyl2025-09-17 | Learn the complete professional workflow for building stunning user interfaces using AI agents like GPT-5 Codex, Claude Code, and Cursor. This comprehensive tutorial covers essential documentation strategies, systematic agent prompting techniques, and step-by-step UI development using modern tools like ShadCN and Next.js. Discover how to gather design inspiration, structure your project for success, and avoid common mistakes that lead to chaotic development. Whether you're using Cursor, Claude Code, or any other AI coding agent, you'll learn the proven methodology that separates professional developers from amateurs. Includes practical examples of building a real application interface, integrating component libraries, implementing version control best practices, and creating maintainable code that scales. Perfect for developers who want to leverage AI agents effectively while maintaining professional development standards and producing production-ready results.
π οΈ TOOLS COVERED:
Firecrawl (reliable web scraping service) 11Labs v3 (advanced text-to-speech with multi-speaker dialogue) Vercel AI SDK (model-agnostic AI integration) Agentic coding using Claude Code, Codex, Cursor, and more.
β° DETAILED TIMESTAMPS: 00:00 - Introduction 01:12 - Documentation setup 04:52 - Configuring coding agents and memory files 06:14 - UI planning with inspiration screenshots 08:15 - Building initial three column layout 09:06 - Creating development checkpoints 09:39 - Installing ShadCN component library 11:00 - Setting up agent rules and constraints 11:59 - Converting to professional ShadCN components 13:11 - Custom theming with TweakCN
Learn professional development practices that coding agents struggle with - proper planning, architecture design, and model flexibility that scales to enterprise applications.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#claudecode #codex #vibecodingI Built an AI Podcast Generator and Heres How You Can TooLeon van Zyl2025-09-16 | Build a complete AI podcast generator that turns articles into engaging conversations! This series covers everything: web scraping, AI conversations, professional UI design, and voice generation using OpenAI, 11Labs, and Next.js. Perfect for developers and no-code builders. Source code included!
In this series, we'll use the new OpenAI GPT-5-Codex model and the Codex CLI to build the app. Other coding agents and CLI tools will work too!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Complete AI podcast generator that scrapes any URL and creates natural dialogue πΉ Professional project architecture and planning techniques for real-world applications πΉ Model hot-swapping strategies to easily switch between AI providers πΉ Next.js full-stack development with proper tech stack decisions πΉ Integration of multiple AI services (web scraping, LLMs, and text-to-speech)
π― PERFECT FOR:
β Developers wanting to build production-ready AI applications β "Vibe coders" looking to level up their development skills β Anyone interested in AI automation and podcast creation β Next.js beginners ready for a comprehensive project β Developers tired of coding agents producing poor results with simple prompts
π οΈ TOOLS COVERED:
Firecrawl (reliable web scraping service) 11Labs v3 (advanced text-to-speech with multi-speaker dialogue) Vercel AI SDK (model-agnostic AI integration) Agentic coding using Claude Code, Codex, Cursor, and more.
β° DETAILED TIMESTAMPS:
00:00 - AI Podcast Generator Demo & Project Overview 01:25 - Planning Phase: Breaking Down Core Functionality 02:27 - Architecture Design: Scrape β Generate β Audio Pipeline 04:02 - Web Scraping Tools: Why Firecrawl Over Alternatives 05:33 - 11Labs v3 Dialogue Demo: Multi-Speaker Conversations 07:02 - Tech Stack Decision: Next.js Full-Stack Framework 08:43 - AI SDK Strategy: Model Hot-Swapping vs Direct APIs 11:01 - Next.js Project Setup & Installation Process 12:28 - Development Server Launch & Series Preview
Learn professional development practices that coding agents struggle with - proper planning, architecture design, and model flexibility that scales to enterprise applications.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#claudecode #codex #vibecodingHow to Use Nano Banana to Build a Complete SaaS App (Step-by-Step)Leon van Zyl2025-09-08 | Build a complete AI-powered interior design SaaS app using Google's breakthrough Nano Banana image editing model - no advanced coding required!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Complete room design app with user authentication and credit system πΉ Integration with Google's Nano Banana AI for professional image editing πΉ Full-stack Next.js application with PostgreSQL database πΉ Production-ready SaaS with modern UI components and themes
π― PERFECT FOR:
β Developers wanting to leverage cutting-edge AI image editing β Entrepreneurs building AI-powered SaaS products β No-code enthusiasts ready to learn modern development β Anyone interested in Google's latest AI capabilities β Freelancers looking to create client applications
00:00 - Nano Banana introduction and Room GPT demo 02:30 - Setting up the development environment 04:00 - Docker PostgreSQL database configuration 06:00 - Google OAuth setup walkthrough 09:00 - Database migration and schema setup 10:00 - Claude Code application building process 17:00 - Testing image editing functionality 19:30 - Adding loading states and user experience improvements 22:00 - Professional UI design system implementation 24:00 - Advanced theme customization with TweakCN 27:00 - Final application showcase and next steps
Google's Nano Banana delivers professional-grade image editing capabilities through simple API calls, making it perfect for building scalable SaaS applications without complex image processing infrastructure.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#NanoBanana #ClaudeCodeHow to Use Claude Code From ANYWHERE (Github Tutorial)Leon van Zyl2025-09-01 | Deploy and modify your applications from anywhere using Claude Code - no computer or code editor required! Make changes to production apps using just your phone and GitHub issues.
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Deploy your application from GitHub to Vercel with automated deployments πΉ Set up Claude Code to automatically implement changes via GitHub issues πΉ Configure environment variables and OAuth authentication for production πΉ Create AI-powered development workflows accessible from mobile devices πΉ Implement automated code reviews and pull request management
π― PERFECT FOR:
β Developers wanting flexibility to make quick changes remotely β Business owners who need rapid application updates on-the-go β Teams looking to streamline their development workflow with AI β Anyone tired of booting up code editors for small changes β Claude Code users ready to take their automation to the next level
π οΈ TOOLS COVERED:
Claude Code (AI-powered coding assistant) GitHub (code repository and issue management) Vercel (deployment platform with automatic deployments)
β° DETAILED TIMESTAMPS:
00:00 - Introduction: Remote development with Claude Code 01:14 - Setting up GitHub repository for your project 02:30 - Deploying to Vercel with environment variables 04:09 - Configuring Google Cloud Platform OAuth settings 06:30 - Installing Claude Code GitHub app integration 07:58 - Creating GitHub issues to request code changes 09:09 - Claude Code automatically implementing changes 10:20 - AI code review process and quality checks 11:28 - Testing changes and merging to production 12:40 - Final results and deployment verification
This workflow gives you the ultimate flexibility to make production changes from anywhere in the world, using nothing but a mobile device and GitHub's web interface.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#ClaudeCode #aiagentsClaude Code + Shadcn/ui = Professional AppsLeon van Zyl2025-08-29 | Watch Claude Code automatically transform basic apps into professional-grade applications using shadcn/ui components! No manual coding required - just AI-powered development automation.
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Replace static loading text with animated skeleton components πΉ Install and configure shadcn/ui component library with AI agents πΉ Set up shadcn MCP server for automated component management πΉ Create responsive loading states that match your page content πΉ Build an FAQ section using accordion components via MCP server
π― PERFECT FOR:
β React/Next.js developers wanting better UX β Anyone using AI agents for web development
π οΈ TOOLS COVERED:
shadcn/ui (Popular React component library) Skeleton Component (Animated loading placeholders) Claude Code (Development automation) shadcn MCP Server (Component installation automation)
β° DETAILED TIMESTAMPS:
00:00 - Loading state problems demonstration 00:41 - Skeleton loaders solution introduction 00:56 - shadcn/ui component library overview 02:05 - Implementing skeleton loaders with AI agent 02:40 - Installing skeleton component properly 03:28 - Testing skeleton loaders across pages 04:02 - shadcn MCP server introduction 04:33 - Installing and configuring MCP server 05:07 - Testing MCP server with accordion component 05:50 - FAQ section results and accordion demo 06:00 - Series recap and next steps
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#claudecode #shadcnClaude Code + Playwright MCP = Perfect UI Every TimeLeon van Zyl2025-08-28 | Stop struggling with UI issues in your Claude Code projects - learn how to give your AI coding assistant browser vision to fix problems automatically!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Set up Playwright MCP server in Claude Code for browser automation πΉ Automatically detect and fix UI clipping issues in Claude Code apps πΉ Create comprehensive automated testing workflows using Claude Code πΉ Troubleshoot visual bugs without manual screenshots in Claude Code πΉ Build a complete data structuring application using Claude Code + my Agentic Coding Starter Kit
π― PERFECT FOR:
β Claude Code users struggling with UI issues in generated apps β Developers new to Claude Code MCP server integrations β Anyone building apps with Claude Code, Bolt, or Lovable platforms β Beginners learning Claude Code advanced workflows and automation β Users of the Agentic Coding Starter Kit + Claude Code combination
π οΈ TOOLS COVERED:
Claude Code (AI coding assistant with MCP browser automation) Playwright MCP Server (browser vision for Claude Code agents) Agentic Coding Starter Kit (free Claude Code boilerplate)
β° DETAILED TIMESTAMPS:
0:00 - Introduction: Common UI issues in Claude Code applications 0:35 - Demo: Application built with Claude Code + Agentic Starter Kit 1:17 - Text clipping in Claude Code generated UI 1:52 - Why Claude Code agents code blind: The UI design challenge 2:07 - Playwright MCP server setup for Claude Code 2:30 - Installing MCP server and connecting to Claude Code 2:47 - Agent troubleshooting UI issues automatically 3:16 - Claude Code agent fixing text clipping in real-time 3:41 - Automated workflow setup 4:15 - Full user journey testing 5:18 - Additional Claude Code bug fixes: Modal overlay resolution 5:53 - Wrap up
Unlike manual screenshot debugging, Playwright MCP gives your Claude Code agent real-time browser vision to identify, fix, and test UI issues automatically - making Claude Code development dramatically more efficient.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#ClaudeCode #PlaywrightMCPBuild ANYTHING with DeepAgent (Complete Tutorial)Leon van Zyl2025-08-26 | Deep Agent by Abacus AI can build complete web apps, create professional presentations, and automate complex research tasks - all through a single AI agent with its own virtual computer!
πΉ Build a fully functional Sudoku game with difficulty levels, hints, and save/load features πΉ Create professional 15+ page technical reports with automated research and citations πΉ Generate PowerPoint presentations with AI-generated visuals and comprehensive content πΉ Use virtual computer capabilities for web scraping and automated flight searches πΉ Master Deep Agent workflows for complex multi-step automation projects
π― PERFECT FOR:
β Developers wanting to automate app creation and testing β Business professionals needing automated research and reporting β Content creators looking for presentation automation β Anyone interested in cutting-edge AI agent technology β Teams seeking efficient document and app generation workflows
β° DETAILED TIMESTAMPS:
0:00 - Deep Agent introduction and capabilities overview 1:30 - Chat LLM platform tour and pricing breakdown 2:16 - Project 1 Building Sudoku game with specifications 4:00 - Testing completed Sudoku game functionality 5:00 - Virtual computer demonstration and explanation 6:00 - Project 2 Technical report creation on MCP 7:30 - Project 3 PowerPoint presentation generation with AI visuals 8:30 - Project 4 Flight search automation demo
Unlike other AI coding tools, Deep Agent can actually test and troubleshoot applications using its virtual computer with vision capabilities, ensuring your projects work perfectly before completion.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#DeepAgent #AIAutomation #ChatLLMClaude Code Can Build N8N Workflows For You!Leon van Zyl2025-08-22 | Build complete N8N workflows using Claude Code without ever touching the canvas - two powerful MCP servers make this possible!
π° SAVE WITH HOSTINGER: Get N8N hosting for only $7/month + 10% off with code "LEON" Link: hostinger.com/leon
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Create AI chatbots with memory and web search using voice commands πΉ Build workflows through direct N8N API integration πΉ Automate workflow creation using browser automation πΉ Test and modify existing workflows without manual clicking πΉ Set up MCP servers for seamless Claude Code integration
π― PERFECT FOR:
β N8N users wanting to speed up workflow creation β Claude Code beginners exploring automation possibilities β Developers interested in AI-powered workflow building β Business owners looking to streamline N8N development β Anyone tired of manually dragging and dropping nodes
π οΈ TOOLS COVERED:
Claude Code (AI coding assistant) N8N MCP Server (direct API workflow creation) Playwright MCP Server (browser automation) N8N (workflow automation platform)
β° DETAILED TIMESTAMPS:
00:00 - Introduction to Claude Code + N8N automation 01:08 - N8N MCP server setup and configuration 02:15 - API key creation and connection setup 03:33 - First workflow creation attempt (basic results) 06:00 - Improved prompt engineering for better results 07:30 - Testing workflow modifications and system messages 09:30 - Hostinger sponsorship: Cheaper N8N hosting solution 11:30 - Playwright MCP server setup and browser automation 12:30 - Live browser workflow creation demonstration 14:30 - Advanced testing: Memory and web search validation 16:00 - Workflow testing automation and final thoughts
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#ClaudeCode #n8nClaude Code Starter Kit: Build Real Apps in MinutesLeon van Zyl2025-08-18 | Build production-ready apps with authentication, databases, and AI in minutes using this powerful Claude Code template - no complex setup required!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Set up a complete full-stack boilerplate with authentication, database, and AI integration πΉ Create an app with Google OAuth, Postgres database, and AI categorization πΉ Deploy your application to production with custom domains πΉ Master version control and checkpointing for AI agent development πΉ Implement custom design systems and UI components automatically
π― PERFECT FOR:
β Beginners who want to build real applications ("vibe coders") β Experienced developers looking to speed up project setup β Anyone wanting to integrate AI into full-stack applications β Developers switching from platforms like Bolt or Lovable β Teams needing production-ready templates with authentication
π οΈ TOOLS COVERED:
Claude Code (AI coding agent) Cursor (alternative AI editors) Next.js (React framework) Postgres Database (via Vercel/Neon) Google OAuth (authentication system) OpenAI API (AI integration) Vercel (hosting and deployment) Git (version control and checkpointing)
β° DETAILED TIMESTAMPS:
00:00 - Introduction to the free boilerplate template 00:49 - Download and initial project setup 02:58 - Database configuration and migration 07:26 - Installing prerequisites (Node.js, Git) 08:16 - Setting up Vercel Postgres database 09:08 - Cloning repository and dependency installation 11:57 - Configuring environment variables 14:05 - Google OAuth setup and client configuration 17:33 - OpenAI API integration setup 19:00 - Building the todo application with Claude Code 26:26 - Creating custom design systems 29:58 - Checkpointing and version control with Git 33:33 - Production deployment to Vercel
Eliminates the frustration of broken backend systems that plague platforms like Bolt and Lovable. Get authentication, database persistence, and AI features working out of the box instead of spending hours on setup.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#ClaudeCode #AIAutomation #FullStackDevelopmentBuild AI Agents with GPT-5 (No Code Required!)Leon van Zyl2025-08-13 | π€ Build powerful AI agents in Flowise with GPT-5 and OpenAI's new Responses API - no complex tool setup required!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Create a multi-modal AI agent with built-in web search, code interpreter, and image generation πΉ Set up GPT-5 in Flowise with reasoning capabilities and native tool access πΉ Generate interactive data visualizations from CSV files using code interpreter πΉ Analyze and transform images with AI-powered generation tools πΉ Deploy and share your AI agent as a public chatbot interface
π― PERFECT FOR:
β Flowise users wanting to leverage the latest GPT-5 capabilities β AI automation enthusiasts looking for no-code agent building β Business owners needing data analysis and visualization tools β Developers exploring multimodal AI applications β Anyone wanting to build agents without manual tool configuration
00:00 - GPT 5 and Responses API introduction 00:57 - Setting up Flowise agent workflow 01:47 - Understanding the Responses API vs Completions API 02:30 - Configuring GPT 5 model settings and reasoning 03:35 - Testing built in web search capabilities 04:28 - Image analysis and AI powered generation demos 06:42 - Enabling file uploads for code interpreter 07:24 - Creating interactive charts from CSV data 09:16 - Sharing your agent as a public chatbot
The combination eliminates the need to manually configure web search, code execution, and image tools - everything works out of the box with OpenAI's native implementations optimized for their models.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#flowise #gpt5 #aiagentsI Used Claude Code as an MCP Server (This Actually Works!)Leon van Zyl2025-08-12 | π€ What happens when you let GPT-5 control Claude Code's tools? A silly experiment that actually works!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Turn Claude Code into an MCP server that any AI model can control πΉ Connect GPT-5 to Claude Code's file editing, bash commands, and directory tools πΉ Build a React to-do app using GPT-5 brain + Claude Code hands πΉ Set up FlowiseAI as an MCP client for cross-model tool sharing πΉ Discover why (or why not) you'd want other AIs controlling Claude Code
π― PERFECT FOR:
β Developers curious about Claude Code MCP server capabilities β People wanting to use GPT-5 for actual coding projects β FlowiseAI users looking for new tool integrations β Experimenters who love combining different AI models
π οΈ TOOLS COVERED:
Claude Code (running as MCP server) FlowiseAI (MCP client and agent builder) GPT-5 (OpenAI's new reasoning model) Model Context Protocol (tool exposure standard)
β° DETAILED TIMESTAMPS:
00:00 - Claude Code MCP server discovery 00:40 - Setting up FlowiseAI as MCP client 01:30 - Configuring GPT 5 agent with Claude Code tools 02:40 - Adding MCP server configuration 04:00 - Testing tool connections and permissions 05:00 - GPT 5 creates project folders using Claude Code 06:00 - Building React app with cross-model collaboration 07:30 - Testing the completed to do list application 08:40 - GPT 5 applies wild neon styling transformations
Expose Claude Code's powerful development tools to any AI model - because sometimes you want GPT-5's reasoning with Claude Code's execution capabilities!
π Subscribe for AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#gpt5 #aiagents #claudecodeRun OpenAIs Open Source Model FREE in n8n (Complete Setup Guide)Leon van Zyl2025-08-06 | π€ OpenAI just released their first open weight model GPT-OSS and it's comparable to frontier reasoning models like o4-mini and o3-mini - run it locally for FREE with zero API costs!
β‘ WHAT YOU'LL BUILD/LEARN: πΉ Set up OpenAI's GPT-OSS model locally using Ollama πΉ Configure free cloud alternative via Grok's infrastructure πΉ Integrate both options seamlessly into N8N workflows πΉ Build AI agents with custom knowledge bases and tool calling πΉ Test reasoning capabilities against paid OpenAI models
π― PERFECT FOR: β Beginners wanting to try advanced AI models without costs β Business owners seeking budget-friendly automation solutions β N8N users looking for OpenAI alternatives β Developers wanting local AI model deployment β Anyone tired of expensive API bills from paid services
π οΈ TOOLS COVERED:
GPT-OSS (OpenAI's new open-source model with Apache 2.0 license) Ollama (Local AI model deployment and management) N8N (Visual workflow automation platform) Groq (Free cloud-based AI model inference)
β° DETAILED TIMESTAMPS: 00:00 - OpenAI GPT-OSS model introduction and capabilities 01:45 - Installing Ollama for local model deployment 02:30 - Downloading GPT-OSS models (20B vs 120B parameters) 04:00 - Setting up Ollama credentials in N8N 05:00 - Testing local GPT-OSS model with knowledge base 06:30 - Analyzing local model performance and limitations 08:00 - Configuring Groq cloud alternative setup 09:00 - Creating Groq API key and N8N integration 09:30 - Comparing Groq vs local model performance 11:00 - Course promotion and final thoughts
π‘ WHY GPT-OSS? Finally get OpenAI-level performance without the API costs! Perfect for experimentation, learning, and budget-conscious automation projects where response time isn't critical.
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#openai #N8N #AIAutomationClaude Code Agents: Build Agent Teams That WorkLeon van Zyl2025-08-05 | Create specialized agents in Claude Code with subagents - no more messy contexts or poor results from cramming everything into one conversation!
β‘ WHAT YOU'LL BUILD/LEARN: πΉ Set up specialized subagents for code review, backend, and frontend tasks πΉ Create clean, focused conversations with dedicated agent contexts πΉ Orchestrate complex workflows between multiple AI agents automatically πΉ Build better apps with agents that have specific system prompts and tools πΉ Master the @ symbol syntax for calling specific agents directly
π― PERFECT FOR: β Claude Code users struggling with complex, messy conversations β Developers wanting specialized AI assistance for different project areas β Teams looking to streamline their AI-assisted development workflow β Anyone building full-stack applications with AI assistance
β° DETAILED TIMESTAMPS: 00:00 - Introduction to Claude Code sub-agents 02:38 - Managing agents with /agents command 02:48 - Creating your first specialized agent (code reviewer) 04:24 - Configuring agent tools and permissions 05:30 - Selecting models and agent colors for identification 06:57 - Testing the code reviewer agent with transaction code 08:02 - Building a Next.js backend engineer agent 09:25 - Creating a frontend specialist with React/Tailwind focus 11:33 - Using @ symbol to call specific agents directly 12:15 - Orchestrating complex workflows between multiple agents 13:24 - Reviewing the improved UI with animations and icons
π GET STARTED: πΊ Previous Video: Building Budget App from Scratch with Claude Code youtu.be/1JDVrQr2pPc
π‘ WHY SUBAGENTS? Sub-agents keep your conversations clean and focused while delivering better results through specialized system prompts and dedicated contexts for each development task.
π CONNECT: π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#ClaudeCode #AIAgents #AnthropicHow To Use Claude Code (Claude Code Tutorial)Leon van Zyl2025-08-04 | How to use Claude Code for BEGINNERS.
Master Claude Code with a complete workflow that works every single time - no coding experience required!
β‘ WHAT YOU'LL BUILD/LEARN:
πΉ Build a complete budget tracking app from scratch using Claude Code πΉ Master essential Claude Code workflows and best practices πΉ Create custom memory files and slash commands for better productivity πΉ Set up automated hooks and audio notifications for long processes πΉ Use planning mode to create technical implementation documents πΉ Implement version control with commits and rollbacks
π― PERFECT FOR:
β Claude Code beginners wanting to learn proper workflows β Developers looking to maximize AI coding productivity β Anyone wanting to build full-stack apps without manual coding
π οΈ TOOLS COVERED:
Claude Code (AI-powered coding assistant) Next.js (Full-stack React framework) SQLite (Lightweight database solution) Drizzle ORM (Database connection layer) Terminal Setup (Command line optimization) Git Version Control (Project checkpoints) Custom Hooks (Automation triggers) Cursor (Code editor integration)
β° DETAILED TIMESTAMPS:
00:00 - Introduction to Claude Code best practices 00:20 - Starting Claude Code and permission modes 01:06 - Terminal setup and workspace optimization 02:06 - Setting up Next.js project structure 03:14 - Adding SQLite database and Drizzle ORM 04:14 - Understanding Claude Code's todo lists 05:02 - Running the development server 05:41 - Creating Claude.md memory files 07:01 - Adding custom memory rules with hashtags 08:12 - Using /init command for automatic setup 09:16 - Building the budget tracking application 11:18 - Fixing styling and error handling issues 13:39 - Creating commits and version control 15:43 - Context management with /clear and /compact 17:23 - Creating custom slash commands 19:43 - Planning mode vs auto-accept modes 21:12 - Creating PRDs and implementation plans 24:02 - Implementing technical specifications 26:00 - Setting up custom hooks for notifications 29:00 - Processing images for styling themes
Claude Code eliminates the need to manually write code while giving you full control over your project structure and implementation decisions, making it perfect for rapid prototyping and production-ready applications.
π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#ClaudeCode #AIAutomation #NoCodeStop Sending Claude Code Prompts Too Early! (Line Break Fix)Leon van Zyl2025-07-29 | Struggling with Claude Code sending your prompts too early? Learn the simple /terminal-setup trick to add multiple lines to your prompts using Shift+Enter. Perfect for complex coding requests!
#claudecode #anthropicn8n Tutorial for Beginners: Complete AI Automation Guide (Build Anything!)Leon van Zyl2025-07-25 | π€ Build powerful AI automation workflows with n8n - no coding required! This complete 98-minute masterclass shows you how to create AI-powered business automation from scratch.
π° **SAVE 10% ON N8N HOSTING**: Get additional 10% off Hostinger! Link: hostinger.com/leon | Code: LEON
β‘ **WHAT YOU'LL BUILD:** πΉ AI-powered email automation system πΉ Smart customer support chatbot with sentiment analysis πΉ Website & Telegram integration πΉ Automated Google Sheets workflows πΉ Production-ready deployments
π― **PERFECT FOR:** β Business owners wanting to automate workflows β Complete beginners (no coding experience needed) β Zapier users looking for more powerful alternatives β Anyone interested in AI automation
β° **DETAILED TIMESTAMPS:** 00:00 - Introduction to N8N and AI Workflow Automation 00:53 - N8N Setup Options (Cloud, Self-Hosted, Local) 11:32 - AI Powered Motivational Quotes Workflow 16:33 - Setting Up AI Models (OpenAI, Anthropic, Ollama) 23:33 - Email Automation and Google Sheets Integration 41:16 - Advanced AI Customer Feedback System 46:16 - AI Sentiment Analysis and Conditional Workflows 56:33 - Building Your First AI Agent with Custom Knowledge Base 1:03:16 - Agent Reservation System Integration 1:09:16 - Agent Human Support Handoff Integration 1:13:33 - Setting Up Persistent Databases (PostgreSQL & Vector Storage) 1:17:00 - Deploying Workflows to Production 1:25:00 - Website Chat Widget Integration 1:30:33 - Creating Telegram Bots with BotFather 1:36:00 - Final Testing and Conclusions
π‘ **WHY n8n?** Unlike Zapier's limited free tier, n8n offers unlimited workflows and runs when self-hosted. Perfect for businesses wanting full control over their automation without monthly limits.
π **CONNECT:** π Subscribe for weekly AI automation tutorials π¦ Follow on Twitter: https://x.com/leonvz
#n8n #AIAutomation #aiagentsClaude Code MCP: How to Add MCP Servers (Complete Guide)Leon van Zyl2025-07-21 | Learn how to supercharge your Claude Code agents by adding MCP servers in this complete tutorial. Discover how to integrate Model Context Protocol servers like Context7 for up-to-date documentation and Bright Data for web scraping capabilities. This step-by-step guide covers both remote and STDIO server setup, environment variable configuration, Windows-specific installation without WSL, and practical demos showing your AI agents accessing live data and APIs. Perfect for developers looking to extend Claude Code functionality with custom tools and integrations.
π TIMESTAMPS 00:00 - Introduction to MCP Servers for Claude Code 00:42 - Types of MCP Servers (Remote vs Local) 02:42 - Adding Remote MCP Server (Context7) 03:56 - MCP Add Command Structure 07:46 - Adding STDIO Servers 10:26 - Environment Variables & API Keys 11:16 - Windows Setup Without WSL 12:20 - Advanced Setup: Bright Data MCP Server 16:09 - Live Demo: Testing Both MCP Servers
#claudecodeClaude Code on Windows - EASY Setup & Cursor Integration (Step-by-Step)Leon van Zyl2025-07-14 | Setup Claude Code on Windows and add to Cursor.
Learn how to set up Claude Code on Windows with native support - no WSL required! This tutorial covers complete installation, basic usage for coding tasks, and integration with popular IDEs like Cursor and VS Code. Perfect for developers wanting to use Anthropic's AI coding agent on Windows. Includes Node.js setup, authentication, and live coding demonstrations.
π TIMESTAMPS: 0:00 - Introduction to Claude Code Windows Native Support 0:25 - Installing Node.js Prerequisites 1:00 - Claude Account Setup (Pro/Max vs API) 2:12 - Installing Claude Code via NPM 2:28 - Authentication and Initial Setup 3:09 - Using Claude Code in Terminal/Command Prompt 4:27 - Building Snake Game Demo with Virtual Environment 6:01 - Adding Claude Code to IDEs (Cursor/VS Code) 7:03 - Two-Player Snake Game Modification Demo 8:23 - Wrap-up and Final Thoughts
#claudecode #anthropicHow to Add Grok 4 to n8n AI Agents (+Vision and Real-Time Search)Leon van Zyl2025-07-14 | Learn how to integrate Grok 4 with n8n workflows in this complete step-by-step tutorial. We'll show you how to set up Grok 4 API credentials, create AI agents, and compare Grok 4's coding performance against Claude Sonnet through a practical coding challenge. Discover Grok 4's multimodal capabilities including vision processing for image analysis and built-in web search features that provide real-time information with citations.
π TIMESTAMPS: 00:00 - Introduction to Grok 4 and n8n Integration 00:29 - Setting Up x.ai API Account and Credentials 01:44 - Creating n8n Workflow with Grok 4 Agent 02:48 - Grok 4 vs Claude Sonnet Coding Challenge 05:00 - Testing Claude Sonnet App Results 07:15 - Testing Grok 4 App Performance and Fixes 11:26 - Adding Vision Capabilities to Grok 4 Workflows 14:25 - Implementing Built-in Web Search Features 16:08 - Final Results and Conclusions
------------------------------------------------------------------------------------- VISION JSON ------------------------------------------------------------------------------------- { "messages": [ { "role": "user", "content": [ { "type": "text", "text": "What do you see in this image?" }, { "type": "image_url", "image_url": { "url": "images.pexels.com/photos/32507137/pexels-photo-32507137.jpeg", "detail": "high" } } ] } ], "model": "grok-4" }
------------------------------------------------------------------------------------- LIVE SEARCH JSON -------------------------------------------------------------------------------------
{ "messages": [ { "role": "user", "content": "What is the latest news from openai?'" } ], "model": "grok-4", "search_parameters": { "mode": "auto" } }
#n8n #aiagents #grok4Build a DEEP Research Agent That Doesnt Suck (Flowise AI Tutorial)Leon van Zyl2025-07-09 | This video demonstrates how to build a powerful deep research agent flow using Flowise that actually works effectively. Learn to create a multi-agent system with a planner orchestrator that spawns specialized subagents for targeted research tasks, uses iteration nodes to process multiple research queries simultaneously, and generates comprehensive reports with proper citations. The tutorial covers integrating web search tools, web scrapers, and research databases while implementing conditional logic to determine when additional research is needed, creating a robust automated research workflow that produces detailed, well-sourced reports.
π TIMESTAMPS: 00:00 - Introduction to Deep Research Agents 00:38 - Problems with Existing Research Models 01:33 - Anthropic's Multi-Agent Research System Theory 02:33 - FlowWise Implementation Overview 05:42 - Setting Up Form Input and Flow State Variables 07:28 - Iteration Node Demo and How It Works 11:00 - Building the Planner Agent with JSON Output 13:00 - Creating Research Subagents with Tools 14:00 - Adding Tavily API and Web Scraper Tools 16:00 - Testing Individual Subagent Performance 17:00 - Building the Writer Agent for Report Generation 19:00 - Adding Condition Agent for Quality Control 20:33 - Implementing Loop Back Logic to Planner 22:00 - Final End-to-End Workflow Testing 23:30 - Results and Conclusion
#flowise #aiagents #deepresearchBuild AI Agents With Human Feedback (Flowise Tutorial)Leon van Zyl2025-07-07 | Learn how to add Human in the Loop functionality to your Flowise AI agent workflows to get better results, reduce token costs, and maintain control over your AI agents. This comprehensive tutorial shows you how to build feedback loops, tool approval systems, and persistent state management that survives server restarts.
π TIMESTAMPS: 00:00 - Introduction to Human in the Loop 00:38 - What is Human in the Loop? 02:38 - Building the Research Agent in Flowise 05:42 - Adding Tools (Google Search & Web Scraper) 07:23 - Testing the Research Agent with Tool Approval 09:00 - Creating Title and Outline LLM Node 11:00 - Adding Human Input Node for Feedback 12:47 - Testing Feedback Loop and Rejections 14:37 - State Persistence Feature (Server Restart Demo) 16:18 - Adding Blog Post Writer Feedback Loop 17:33 - Setting Up Email Agent with Gmail Integration 19:16 - Google Cloud Console OAuth Setup 22:20 - End-to-End Workflow Testing 23:30 - Final Results and Conclusion
#flowise #aiagentsHow I Add REAL-TIME Web Scraping to Any AI Agent (Bright Data MCP)Leon van Zyl2025-06-25 | Learn how to add powerful real-time web scraping capabilities to your AI agents using Bright Data's MCP server in this comprehensive Flowise tutorial. Discover why traditional search tools like Brave Search and Serp API only perform basic Google searches without actually scraping website content, and see how to overcome web scraping challenges like IP blocking, captchas, and dynamic content. This step-by-step guide shows you how to set up the MCP server, configure API keys, and enable your AI agents to extract real-time data from any website including Amazon product searches, OpenAI news articles, and hundreds of other platforms. Perfect for developers looking to enhance their AI workflows with limitless web data infrastructure that bypasses all traditional scraping limitations.
π TIMESTAMPS: 0:00 - Intro 0:22 - Limitations of Traditional Search Tools 2:18 - Web Scraping Challenges 3:24 - Setting Up Bright Data Account 4:07 - Configuring MCP Server in Flowwise 6:01 - Testing OpenAI News Scraping 7:01 - Amazon Product Search Demo 8:28 - Results
#flowise #aiagents #webscrapingHow to Add MCP Servers to AI Agents (FlowiseAI Tutorial)Leon van Zyl2025-06-21 | Learn how to supercharge your Flowise agents by adding MCP (Model Context Protocol) servers for real-time documentation access. This step-by-step tutorial shows you how to upgrade from basic LLM nodes to powerful agent nodes that can call tools and retrieve up-to-date documentation using Context7 MCP integration. Follow along as we enhance a software development AI team and build a Flowise assistant that accesses the latest Next.js, Tailwind, and Flowise documentation in real-time. Perfect for developers looking to build smarter AI agents that provide accurate, current information instead of outdated responses. Includes complete setup instructions for custom MCP server configuration, tool calling demonstrations, and practical examples you can implement immediately in your own Flowise workflows.
π TIMESTAMPS: 00:00 - Introduction & MCP Enhancement Request 00:42 - What is MCP (Model Context Protocol)? 01:34 - Context7 Documentation Repository Overview 01:49 - Replacing LLM Node with Agent Node 03:23 - Adding Custom MCP Server to FlowWise 04:47 - Testing Enhanced Software Development Team 05:15 - Analyzing MCP Tool Usage & Results 06:48 - Building Flowise Assistant Example 07:56 - Testing Flowise Assistant with Real Documentation
#flowise #aiagents #mcpserverHow to Build AI Agent Teams in Flowise (Step-by-Step)Leon van Zyl2025-06-20 | Learn how to build powerful AI teams in Flowise with this comprehensive step-by-step tutorial covering supervisor patterns, multi-agent workflows, and intelligent task coordination. This guide shows you how to create a supervisor system that orchestrates specialized AI workersβincluding software engineers and code reviewersβthrough conditional routing, flow state management, and JSON structured output. You'll master advanced Flowise techniques like enum validation, loop nodes, and worker coordination to build production-ready AI systems capable of handling complex development tasks autonomously. Perfect for developers wanting to move beyond single AI implementations, this tutorial demonstrates practical team orchestration using GPT-4, smart routing patterns, and collaborative workflows that actually work together to deliver complete solutions.
π TIMESTAMPS: 00:00 - Introduction to Supervisor Teams in Flowise 00:41 - Creating the Supervisor Node Setup 02:20 - Adding Conditional Routing Logic 03:23 - JSON Structured Output & Enum Values 05:42 - Flow State Management Implementation 09:04 - Building Worker Nodes (Software Engineer) 11:47 - Adding Code Reviewer Worker 14:22 - Final Answer Generator & Testing 16:41 - Conclusion and Next Steps
#flowise #aiagentsShared State in Flowise is Literally Game ChangingLeon van Zyl2025-06-19 | Learn how to use Flowise shared state to build more efficient AI agent workflows that outperform traditional conversation history approaches. This complete tutorial covers Flow State variables, conditional logic, and real examples showing how to optimize token usage and improve LLM performance in your Flowise projects.
π§ I can build your chatbots for you! cognaitiv.ai
π TIMESTAMPS: 00:00 - What is Shared State in Flowwise 01:57 - Problems with Conversation History 03:03 - Introduction to FlowState Variables 05:08 - Building Your First Shared State Flow 12:31 - Setting Up FlowState Variables 16:40 - Advanced Conditional Logic Example 21:08 - Using Custom Functions vs LLM Nodes 26:50 - Final Implementation & Results
#flowise #aiagentsVectorShift Just Launched AI Agents - Heres Everything NewLeon van Zyl2025-06-18 | VectorShift just revolutionized AI agent development! Learn how to build powerful AI agents without coding in this complete tutorial. Watch me create an AI agent that performs web research, queries custom knowledge bases, scrapes websites, sends emails, and generates detailed reports - all automatically. This VectorShift AI agents tutorial covers everything from basic setup to advanced pipeline integration, perfect for anyone wanting to automate workflows with no-code AI solutions. Start building AI agents today!
π TIMESTAMPS: 00:00 - VectorShift AI Agents Introduction 00:26 - Accessing the New Agents Feature 00:55 - Creating Your First AI Agent (Jarvis) 01:20 - Basic Agent Configuration & Setup 01:41 - Setting Agent Instructions & System Prompts 02:02 - Adding Custom Knowledge Base Tool 02:21 - Creating Oak and Barrel Knowledge Base 03:17 - Testing Knowledge Base Integration 03:51 - Adding Web Search Capabilities 04:25 - Testing Real-Time Weather Search 04:42 - Adding URL Scraping Tool 05:16 - Testing Website Content Extraction 05:32 - Adding Gmail Email Integration 06:29 - Testing Complete Workflow (Scrape + Email) 06:52 - Creating Blog Post Generator Pipeline 08:22 - Testing Blog Post Generation 08:42 - Adding Deep Research Pipeline 10:17 - Deploying Agent in Pipeline Workflow 11:00 - Testing Complete Research & Email Automation 12:33 - Viewing Agent Execution Trace 13:16 - Final Results & Wrap-Up
#vectorshift #aiagentsHow to Host AI Agents for $5/Month (Flowise + Hostinger Guide)Leon van Zyl2025-06-16 | Get an additional 10% when you sign up for Hostinger! http://hostinger.com/leon10 Use coupon code LEON at checkout.
Discover how to host your FlowiseAI agents on a VPS for just $5/month instead of $35 with Hostinger. This tutorial shows you step-by-step how to set up a powerful FlowiseAI instance on Hostinger - perfect for building AI agents and advanced agentflows. You'll learn how to configure your server, create a simple AI agent with web search capabilities, and maintain your instance.
π TIMESTAMPS: 00:00 - Why Run Flowise Locally 00:40 - Limitations of Local Setup 01:01 - Flowise Cloud vs Self-Hosting 02:13 - Hostinger VPS Setup for Flowise 03:33 - Installing Flowise on Hostinger 04:34 - Accessing Flowise Dashboard 05:17 - Keeping Flowise Up to Date 06:52 - Building the First Agent 08:00 - Connecting Knowledge Base 09:30 - Setting Up Vector Database 10:57 - Fixing SSL Certificate Error 11:46 - Uploading Documents to Vector DB 12:16 - Linking Knowledge Base to Agent 12:50 - Testing the Agent 13:43 - Embedding Agent on a Website 14:33 - Final Thoughts & Subscribe Call-to-Action
#flowise #aiagents #hostingerHow to Build a Local AI Agent With Flowise (Ollama, Postgres)Leon van Zyl2025-06-13 | In this easy Flowise tutorial, youβll learn how to build powerful AI agents that run completely on your own computer β no cloud, no coding needed. Weβll show you how to set up Flowise, connect it to a local knowledge base using a vector database, and start chatting with your custom data in minutes. Whether youβre using local models or just want full control and privacy, this video will guide you through everything step-by-step. Perfect for creators, educators, and entrepreneurs using no-code tools to build AI workflows with Flowise.
π TIMESTAMPS: 00:00 - Intro 00:10 - Why Run Locally 00:30 - Flowise Local Setup 00:40 - Vector Database Setup 00:45 - Record Manager 01:01 - Running Local Models 02:01 - Knowledge Base Upload 02:30 - Testing Agent 10:17 - Final Thoughts
#flowise #aiagentsSTOP Building Dumb AI Agents β Do This InsteadLeon van Zyl2025-06-09 | Learn how to supercharge your AI agents in Flowise using Retrieval Augmented Generation (RAG). In this tutorial, we explore multiple ways to enhance your agent's ability to answer questions by integrating external tools like the SERP API for live Google searches and custom knowledge bases using document stores. Whether you're building chatbots or internal AI tools, this no-code walkthrough will help you build smarter, context-aware agents that actually understand and respond with relevant information. Perfect for beginners and Flowise v3 users looking to level up their AI workflows.
π TIMESTAMPS 00:00 - Intro 00:02 - Overview of AI Agent from Previous Video 00:35 - Assigning Tools to the Agent 01:14 - Uploading Documents and Images 02:24 - Asking the Agent Questions 03:03 - Customizing the Prompt 04:03 - Adding System Prompt Context 05:09 - Troubleshooting Context Awareness 06:02 - Providing Context Dynamically 07:33 - Agent Performing a Google Search 08:37 - Using the SERP API Tool 09:32 - Context-Based Responses from Agent 10:44 - Uploading a PDF Document 11:35 - Large Document Limitations 12:32 - Retrieving Relevant Documents 13:32 - Uploading Chunks to Vector Store 14:42 - Naming the Knowledge Base 16:02 - Connecting the Document Store 17:00 - Using Agent Flows 18:09 - Setting Up Document Store Tool 19:39 - Defining the Knowledge Base 20:32 - Enabling File Uploads 21:36 - Allowing Image Uploads 22:00 - Wrapping Up
#flowise #aiagents #retrievalaugmentedgenerationYOU WONT BELIEVE How Simple Building AI Agents Gets with Flowise v3Leon van Zyl2025-06-04 | π Learn how to build powerful AI agents using Flowise v3 β a no-code AI builder packed with new features like AgentFlows, custom knowledge bases, web search integration, and email automation. In this step-by-step tutorial, weβll walk you through creating an AI assistant that can search Google, answer questions from uploaded documents, and connect to tools like Gmail and external APIs.
π§ Youβll also discover how to set up document stores, chunk and vectorize your data with OpenAI embeddings, and integrate a Postgres vector database using platforms like Supabase. Plus, we explore real-world tools like SERP API and Composio to extend your AI agentβs capabilities even further.
π TIMESTAMPS 00:00 - Intro & Whatβs New in Flowise v3 00:43 - Demo: Querying a Custom Knowledge Base 01:46 - Performing a Live Web Search with an AI Agent 02:35 - Sending Emails via Gmail Integration 03:30 - Building AI Agents Using Flowise Agentflows 04:28 - Adding Tools like Web Search & Calculators 05:33 - Using SERP API in Flowise 06:56 - Handling Current Date with Date/Time Tool 08:00 - Creating and Managing Document Stores 09:17 - Uploading Data & Chunking Documents 10:44 - Vector Database & Embeddings Setup 12:00 - Avoiding Duplicates with Record Manager 13:27 - Uploading & Abserting Menu Data (CSV) 14:33 - Testing Document Retrieval 15:10 - Using the Document Store in Your Agent 16:00 - Integrating with External Apps using Composio 17:26 - Setting up Gmail Integration via Composio 18:52 - Sending Emails from Agent (Demo) 19:45 - Outro & Whatβs Coming in Future Videos
#flowise #aiagentsFlowise v3 Complete Tutorial: Build AI Agents WITHOUT CodingLeon van Zyl2025-06-03 | Discover everything you need to know about Flowise v3, the open-source no-code platform for building powerful AI-driven solutions. In this video, we cover Agent Flows v2, seamless LLM integration (including OpenAI), and advanced features like human-in-the-loop, custom knowledge bases, and document store functionality. Learn how Flowise puts AI capabilities front and centerβunlike N8N, Zapier, or MEG.comβand find out how to install it locally, use Flowise Cloudβs free tier, and self-host for unlimited flows. We walk through the Flowise dashboard, demonstrate creating your first AI agent, and explore state management, conversation memory, and debugging tools to optimize performance. Whether youβre a developer, no-code enthusiast, or AI hobbyist, this tutorial will help you leverage Flowiseβs robust features to build multi-agent AI systems, automate workflows, and integrate external APIs with ease.
π TIMESTAMPS: 00:00 - Intro to Flowise v3 00:16 - Key Features Overview 00:55 - Flowise vs N8N, Zapier, MEG.com 01:39 - Flowise Open Source & Licensing 02:28 - Flowise Cloud Free Tier & Pricing 03:40 - Installing Flowise Locally 05:00 - Flowise Dashboard Walkthrough 05:45 - Chat Flows vs Agent Flows v2 07:31 - Building AI Agents with Flowise 09:16 - Custom API & Integrations 10:15 - Document Store & Knowledge Bases 11:18 - Creating Your First Agent Flow 13:33 - Adding LLM Nodes (OpenAI) 16:46 - Conversation Memory & State Management 18:33 - Executing & Testing Flows 20:22 - Advanced Node Configurations 22:04 - Debugging & Execution Logs 22:30 - Next Steps & Outro
#flowise #aiagentsBuild an AI Chatbot in Minutes (Vectorshift)Leon van Zyl2025-06-02 | Discover how to use Vectorshiftβs new conversational nodesβTalk and Listenβto build a restaurant chatbot. We cover integrating a custom knowledge base with OpenAI for Q&A and syncing reservations to Google Sheets.
π TIMESTAMPS: 00:00 β Intro & Demo 02:39 β Create Conversational Pipeline 04:24 β Collecting User Details (Name & Email) 07:02 β Knowledge Base Q&A Setup 13:58 β Looping & Merge Node 15:36 β Reservation Flow Setup 16:19 β Google Sheets Integration 18:09 β Final Demo & Outro
#vectorshiftBuild a Powerful AI Image Generator using n8n, Lovable and OpenAILeon van Zyl2025-05-06 | Learn how to create a powerful AI image generator without writing a single line of code! This step-by-step tutorial shows you how to combine n8n workflows with Lovable's AI-powered UI builder and OpenAI's cutting-edge Image-1 model. Perfect for no-code enthusiasts and AI beginners, you'll discover how to set up webhook integrations, configure the OpenAI API, and troubleshoot common issues. The result? A professional, modern application that generates high-quality AI images from text prompts - with exceptional text rendering capabilities that outperform other models. Whether you're building AI solutions for fun or professional use, this accessible tutorial demonstrates how powerful AI technology can be harnessed without programming knowledge. Join the no-code AI revolution and start creating your own image generation applications today!
#n8n #lovable #openaiHow to Host n8n in the Cloud Cheaper (Save 70%)Leon van Zyl2025-04-29 | Get an additional 10% when you sign up for Hostinger! Link: hostinger.com/leon Use Code LEON as checkout.
Discover how to host your n8n workflows in the cloud for just $7/month instead of $25+ with Hostinger. This tutorial shows you step-by-step how to set up a powerful n8n instance on Hostinger - perfect for AI automation projects and client workflows. You'll learn how to configure your server, create a simple AI agent with web search capabilities, and maintain your instance. Stop overpaying for cloud hosting and deploy production-ready n8n workflows today!
π TIMESTAMPS: 00:00 - Intro 01:06 - Hostinger Setup 03:23 - Setting up n8n 05:19 - Accessing n8n 06:18 - Creating an AI Agent 10:59 - Update n8n
#n8n #aiagents #hostingerSave HOURS! Screen Resumes using AI (Vectorshift Tutorial)Leon van Zyl2025-04-22 | Discover how to build an AI-powered resume screening system with VectorShift's no-code platform! This tutorial walks both technical and non-technical professionals through creating a workflow that automatically evaluates candidates against job specifications, providing match percentages and AI reasoning. Perfect for agencies and consultants looking to deliver innovative AI solutions to clients without writing code. Learn to leverage VectorShift's new Workbooks feature while building a practical recruitment tool that saves hours of manual work. Elevate your service offerings with AI automation that impresses clients and delivers immediate ROI. Start building client-ready AI solutions today!
π TIMESTAMPS: 00:00 - Intro 00:36 - Demo 01:04 - Into the VectorShift 01:32 - Create new Workbook 02:34 - Uploading Job Spec 02:50 - Get Candidates from Google Sheet 04:09 - Creating a sub pipeline 05:26 - Adding LLM 08:08 - JSON Output 11:32 - Completing main pipeline 12:23 - List mode 14:09 - Add rows to spreadsheetHow To Connect WhatsApp to N8N (Step-by-Step Tutorial)Leon van Zyl2025-04-16 | Learn how to build a sophisticated WhatsApp AI agent using N8N workflow automation in this step-by-step tutorial. Transform your WhatsApp experience with an intelligent assistant capable of processing text, images, and voice messages while responding with both text and audio replies. This comprehensive guide walks you through the entire process from setting up the Meta Developer platform to creating a fully functional AI-powered WhatsApp bot. The tutorial covers everything you need: configuring the WhatsApp Business Cloud API, building dynamic N8N workflows, implementing OpenAI for voice transcription and image analysis, and creating conditional paths for different message types.
Key features demonstrated include: - Processing and responding to text messages - Transcribing voice messages and generating audio responses - Analyzing images and answering questions about them - Maintaining conversation memory and context - Integrating web search for real-time information - Accessing emails and to-do lists
Perfect for automation enthusiasts, business owners wanting to enhance customer service, or developers looking to build practical AI applications. Download the workflow template for free and start building your own WhatsApp AI agent today!
π TIMESTAMPS: 00:00 - Intro 00:39 - Demo 02:26 - Meta Whatsapp App setup 04:32 - n8n Whatsapp Trigger 06:05 - Adding AI Agent 07:51 - Response to Whatsapp 09:26 - Handling different input types (SWITCH) 13:47 - Audio Input 17:11 - Respond with Audio 21:08 - Image Input 25:45 - Tools 26:28 - Note on Meta VerificationBuild in MCP Server in n8n within minutes!Leon van Zyl2025-04-11 | n8n now officially supports MCP (Model Context Protocol), which means you can create an MCP server in n8n and expose tools to AI agents apps like Cursor and Claude Desktop. You can also call MCP servers as tools from within n8n.
#n8n #aiagentsBuild Anything with MCP in n8n, Heres How!Leon van Zyl2025-04-10 | N8n just added official support for MCP (Model Context Protocol)! This tutorial shows you how to build custom AI tools that connect directly with Claude Desktop and Cursor. Learn to create MCP servers for to-do management, emails, and content generationβall without coding. Discover how to configure your server, integrate AI applications, and even use N8n as an MCP client to access external services. Perfect for automation enthusiasts wanting to enhance workflows with powerful AI capabilities.
π TIMESTAMPS: 00:00 - Intro 01:03 - MPC Server Demo 05:04 - Create MCP Server 06:15 - Add n8n MCP Server to Claude Desktop 07:34 - Add n8n MCP Server to Cursor 08:50 - Calling other workflows 12:00 - To Do List 15:02 - MCP Client Node 16:30 - Zapier SSE
#n8n #aiagentsHow to Build a Local AI Agent With n8n (NO CODE!)Leon van Zyl2025-04-09 | Discover how to build a powerful AI agent that runs completely on your local machine in this step-by-step tutorial. Learn to combine N8N's workflow automation with Ollama's LLM capabilities and PostgreSQL's vector database features to create a responsive assistant that can access your documents, maintain conversation history, and retrieve real-time information from the web.
This video demonstrates a cost-effective solution requiring no paid API services or cloud subscriptions. You'll see how to properly configure Ollama with two essential models, set up Docker with PostgreSQL for vector search, and orchestrate everything through N8N's intuitive interface.
Perfect for developers, AI enthusiasts, and business users looking to implement private AI solutions, this tutorial covers common troubleshooting issues and provides simplified alternatives to more complex setups. Watch as your agent retrieves information from uploaded documents, maintains context across conversations, and delivers accurate responses using locally-run large language models.
If you're interested in self-hosted AI tools that respect your privacy while delivering powerful functionality, this comprehensive guide walks you through building an extensible foundation you can customize for your specific needs.
#n8n #aiagents #ollamaAutomate Your Browser with AI! Build a Computer Using Agent (OpenAI API)Leon van Zyl2025-04-02 | This engaging tutorial demonstrates how to create Computer Using Agents (CUAs) that can automate browser tasks on your behalf. The video walks through implementing an AI agent that can interact with web browsers to perform searches, navigate pages, and click on elements using the OpenAI API and Playwright. Viewers will learn how to set up a browser environment, implement a feedback loop between the agent and browser, handle various actions like clicking, typing, and scrolling, and troubleshoot common issues like managing multiple tabs. Perfect for developers looking to build AI assistants that can interact with web interfaces, this step-by-step guide covers everything from initial setup to executing complex web navigation tasks automatically.
#openai #openaiapi #aiagentsHow to add Custom Tools to AI Agents (EASY!)Leon van Zyl2025-04-01 | Learn how to enhance your AI agents with custom tools using the OpenAI Responses API in this step-by-step Python tutorial. Discover how to build practical functionalities like a to-do list retriever and real-time weather data integration that your AI can dynamically access. This guide walks you through creating function schemas with Pydantic, handling tool calls, and managing conversation context to create more powerful and versatile AI assistants. Perfect for developers looking to extend AI capabilities beyond built-in features and create agents that can interact with external data sources.