Claude Code Unleashed

Summary

For the past two months, I've been obsessively dogfooding background agents while building Terragon. This quickly became an experiment in unleashing Claude Code from the constraints of local development. The breakthrough wasn't just using AI to write code faster. It was realizing what becomes possible when you: Make it trivially easy to kick off new agents Free these agents from your local machine Run an unlimited number of them at the same time I've landed on a hybrid approach and I want to share and document what I've learned. I'm sure this will continue to evolve and change over time. Claude Code & Claude Max If you haven't tried Claude Code yet, you're missing out. Almost every conversation about claude code is either about how good it is or how expensive it is. Last I checked, just saying "Hello" to claude code costs $0.30! Once you start to dig into the cost of claude code, you'll inevitability hear about "Claude Max" - which is Anthropic's subscription so you pay a monthly fee for some amount of usage. If you pay for the $200/month plan you get effectively unlimited usage. There are still rate limits but they are very generous. On a typical day, I'd otherwise pay almost $100 without my claude max subscription. So in a strange way, Claude max is cheap! (Which also happens to be the other flavor of conversation people have about claude code.) I highly recommend getting a claude max subscription. Not having to worry about rate limits, per-use cost, whether you should use sonnet or opus is truly a game changer. With Max, I stopped thinking in terms of "is this a task that's good for claude code? Is this a task for sonnet or opus?" Instead, it became "How can I offload more work to claude code?" That mental shift alone was worth the price. How do I run more than one agent at a time? Once you embrace using claude code, you'll quickly want to run more than one agent at a time. Part of the reason why claude code is so good is that it takes the time to understand the ...