Using Amp

The docs explain what Amp does. This section is about how we use it: how we get a codebase ready for orbs, how we work with agents through the day, and what each of us still does by hand.

Guides

Longer reads from the team on getting a codebase ready for orbs and working with agents.

Orbs, Explained Everything has changed for us with orbs. How do I make you believe me? Thorsten Ball · 25 Aug 2026

Putting an Agent in an Orb How we made our codebase work for agents on headless remote machines. Thorsten Ball · 2 Jul 2026

How to Pair With an Agent Trust isn't a feeling. It's a passing test suite. Harry Charlesworth · 14 Jan 2026

Feedback Loopable Build a playground, run experiments, and make it feedback loopable. Lewis Metcalf · 5 Feb 2026

Pave The Road A call to action, and why the latest model doesn't matter. Alex Kemper · 10 Aug 2026

How We Use Amp

Everyone on the team gets the same 10 questions: where they talk to Amp, which mode they reach for, how they use Orbs and Puck, how they review changes, prompts they keep reusing, and what they still do by hand.