See exactly how descriptions become working software in real workflows -- the prompts, the configs, the corrections, the shipped result. Every workshop comes from production experience.
Video walkthroughs, downloadable prompt templates, and working code. Not theory -- real systems you can express into existence today.
How to use Claude Code, Cursor, and Windsurf to build full applications from natural language descriptions.
Ollama, DeepSeek, Qwen on a Mac mini. Model routing, RAG pipelines, and zero-cloud privacy.
Build and connect MCP servers. Give your AI agents real tools -- browser, email, calendar, databases.
On-device Whisper, push-to-talk hotkeys, and command routing. Build with your voice, fully offline.
From zero to a fully operational AI development environment. Claude Code, Ollama, MCP servers, and your first agent workflow.
Use a Mac mini as a dedicated AI server over Thunderbolt. Ollama, Qdrant, model routing, and SSH automation.
Install, configure, and ship your first feature with Claude Code. CLAUDE.md, hooks, rules, and permission modes.
How I structure CLAUDE.md for maximum agent performance. Rules, autonomy frameworks, tool enforcement, and credential management.
Orchestrate teams of AI agents with task lists, message passing, and parallel execution. Build a shipping squad.
Pre-tool and post-tool hooks that keep agents safe. Budget guards, context overflow prevention, and rate limiters.
Install Ollama, pull your first models, and run inference locally. DeepSeek, Qwen, Llama -- which model for which task.
Qdrant vector database, nomic-embed-text embeddings, and hybrid BM25+semantic search. Index your entire codebase.
Match the right model to the right task. 3B for triage, 14B for extraction, 32B for complex reasoning. VRAM budgeting on 24GB.
What MCP is, how it works, and how to give your AI agents real tools. Browser, file system, memory, and more.
Create your own MCP server from scratch. Tool registration, resource handling, and connecting to any IDE.
The exact MCP configuration I run daily. Chrome, AppleScript, Notion, Tavily, memory -- and why each one matters.
Set up Whisper on-device, configure push-to-talk, and route voice commands to AI agents. Fully offline.
Explore, Plan, Execute, Verify. The workflow pattern that makes AI agents actually reliable and predictable.
Run separate Claude profiles for web, legal, finance, and OSINT. Each with dedicated MCPs, rules, and skills.
Full walkthrough: scaffold a SaaS with Next.js, Stripe billing, auth, and deploy to Vercel. All with AI agents.
From Figma reference to deployed site using AI agents and browser automation. My exact production workflow.
GitHub Actions, automated testing, and deployment pipelines. Let AI agents handle your release process.
Join builders who direct AI agents every day. Full access to all workshops, prompt templates, and the community.