Uploaded January 2026 | Updated September 2026, 3 weeks ago

---

🚀 Access ALL video resources & get personalized help in my community:
skool.com/agentic-labs

🗨️ Speech to Text App (wisprflow): wisprflow.ai/r?WALTER67
🚀 AutoCoder (now AutoForge): github.com/AutoForgeAI/autoforge
☕ Buy me a coffee: buymeacoffee.com/leonvanzyl
💵 Donate using PayPal: paypal.com/ncp/payment/EKRQ8QSGV6CWW

In this video you will learn how long-running agents are transforming agentic coding with Claude Code. Watch as I compare two projects built with the exact same prompt to show why single context window development fails for complex applications. You will learn how the Anthropic Long Running Agent Harness breaks down massive projects into manageable features, implements them one by one with regression testing, and produces production-ready applications on autopilot. I'll also show you how to set up and use my free Agentic Coding Harness UI for your own projects.

⏰ TIMESTAMPS:
00:00 Long running agents introduction
00:35 Claude Code project comparison
01:45 Single context window results
05:07 Context window limitations explained
06:28 Anthropic harness solution overview
07:01 Long running agent results
10:15 Nano Banana thumbnail generation
13:57 Agentic Coding Harness setup
17:30 Creating projects with UI
21:30 Feature list generation process
25:00 MCP tools coding agents
26:00 Browser testing automation
27:20 YOLO mode vs testing

#claudecode #vibecoding