Video Library
Video training, project ride-alongs and guides for building with AI.
All Videos
Topics
Agent-Browser
Agent OS
Agent Teams
Antigravity
Browsers
Builder Stories
Claude.ai
Claude Code
Claude Code Web
Claude Skills
Cloud Agents
Cloudflare
Codex
Codex CLI
Composer
Context
Cursor
Design
Design OS
Documentation
Figma
Frontend
Gemini
Git
GitHub
Google
Image Generation
Implementation
Jobs To Be Done
Markdown
Marketing
Mindset
n8n
New Products
Obsidian
OpenClaw
Opus
PDF
Plan mode
Profiles
React
Ruby on Rails
SaaS
Shortcuts
Sonnet
Spec-Driven Development
Specs
Standards
Stimulus JS
Subagents
Swarms
Tailwind CSS
TDD
Tests
The Builder Method
UI Design
Verification
VS Code
Web design
Workflow
Worktrees
YouTube
No results
Vercel's agent-browser is a CLI-based, headless browser automation tool that's significantly faster and more token-efficient than the Chrome integration — and it can capture screenshots and video as proof that the feature actually works. You'll see it in action building and visually verifying a confetti effect on the to-do list app, producing shareable artifacts that make code review easier.
New
Pro
The Claude Code Chrome integration lets the agent launch your browser and click around like a real user, but it's slow and can be unreliable — especially with screenshots. It's worth knowing about for quick checks, but the next lesson covers a faster, more practical tool for browser testing in your dev workflow.
New
Pro
Without tests, the agent just writes code and re-reads it, saying "looks good to me" — but that doesn't mean it actually works. TDD gives the agent an objective pass/fail signal, turning testing into an alignment strategy that builds confidence into the process alongside your specs.
New
Pro
Agent teams let you spawn multiple Claude Code sessions that share a task list, coordinate through a mailbox system, and work in parallel on complex features. Brian demos the full workflow by having three agents — backend, frontend, and tester — collaborate on building a project dashboard from scratch.
Pro
Brian demonstrates building a custom code review subagent from scratch using Claude Code's agent builder, covering tool selection, model choice, and agent memory configuration. He then runs it against a real feature to show the full workflow from creation to structured review output.
Pro
Subagents give Claude Code the ability to delegate work to separate context windows, keeping your main session clean while getting more done. This lesson covers the built-in subagents, shows delegation in action on a real feature build, and walks through creating your own custom agents.
Pro
Build two custom Claude Code Skills from scratch — a simple one that spins up new projects from a GitHub template, and an advanced one with a bundled script that generates formatted changelogs from your git commit history.
Pro
Walk through using a GitHub repository template to spin up new projects with a pre-configured tech stack — first manually through the dashboard, then with the GitHub CLI command that we'll turn into a reusable Claude Code Skill.
Pro
Understand what Claude Code Skills are, how they're structured, and how the invocation model works — from automatic discovery based on description matching to manual slash command triggers. Learn the essential concepts before building your first custom Skill.
Pro
Claude Code automatically generates a tracked task list with dependencies from your plan or spec, and you can configure it to persist across sessions — so even multi-day projects pick up right where you left off.
Pro
The Ask User Question tool is one of the most underrated features in Claude Code — it flips the dynamic so that Claude interviews you with smart, targeted questions to resolve design and architecture decisions upfront, rather than making assumptions that lead to rework.
Pro
Plan mode is where every new feature should start in Claude Code — it restricts Claude to read-only operations so you can collaboratively explore, ask questions, and produce a detailed written plan before a single file is created or edited.
Pro
Spec-driven development flips the traditional build cycle on its head — instead of prompting and praying, you spend the majority of your time planning collaboratively with your AI agent before a single line of code is written. This lesson introduces the philosophy behind SDD and explains why the planning phase, not the building phase, is where you as a human make your biggest impact.
Pro
The context window is Claude Code's short-term memory — everything Claude can see and work with in the current session. This lesson explains how it fills up, how to monitor it, and practical strategies for keeping your context clean so Claude builds what you actually want.
Pro
Walk through the essential Git workflow — branches, commits, pushes, PRs, and merges — first manually to understand what's happening, then let Claude Code handle it all through simple prompts and a custom slash command that automates the entire process.
Pro
See what it looks like to launch Claude Code for the first time, create a project folder, and start giving it tasks — from understanding the interface and reading diffs to handling permission prompts and toggling edit modes.
Pro
Understand how Claude Code organizes its configuration through the .claude folder structure — from project-level settings and custom commands to system-level folders that make your setup available across every project.
Pro
Find out which Claude plan you actually need to use Claude Code and access the latest frontier model, how the token limits work in practice, and a straightforward recommendation on where to start and when to upgrade.
Pro
Learn what YOLO mode actually does (--dangerously-skip-permissions), the real risks of skipping permissions, and the safety nets that make it a practical choice for daily use. Plus, set up a terminal alias to launch it with a simple shortcut.
Pro
See how Claude Code runs inside the Cursor IDE and why this pairing works so well — from inspecting files and making quick tweaks to flipping between Claude Code and Cursor's agent for different types of tasks.
Pro
Builder Methods Pro
Unlock Full Access
Get unlimited access to the complete video library, exclusive training, and join our community of AI-first builders.
Learn More
