How developers use mrq to protect their work during AI-assisted development.
You're building a side project or working independently. AI tools help you move fast, but when they break something, you're on your own to fix it.
mrq gives you a safety net without the overhead of meticulous git hygiene. Work fast, experiment freely, and restore when needed.
Typical workflow:
mrq watch --daemon
# Code with AI all day
# Restore if something breaks
mrq restore abc123You're in flow, prompting an AI agent to make change after change. The code is evolving rapidly. You don't want to stop and commit every few minutes.
mrq captures state continuously in the background. When the session ends, you can review what changed and commit the result to git with a clean message.
No commit noise during exploration. Clean history when you're done.
You're asking an AI to refactor your authentication system, restructure your API, or modernize a legacy module. These are high-risk changes that touch many files.
With mrq, you can let the AI attempt ambitious refactors. If the result doesn't work, restore to before the refactor started and try a different approach.
Recovery time: seconds, not hours of manual reversion.
You're using AI to help learn a new framework or language. You're following along with suggestions, accepting code you don't fully understand yet.
mrq lets you experiment without fear. Try the AI's suggestions. If they lead somewhere confusing, restore to a point where things made sense.
Explore freely, restore when lost.
You're building a prototype or MVP quickly. The goal is to validate an idea, not maintain perfect code. AI helps you move fast.
Speed matters, but so does not losing work. mrq runs silently, capturing state without slowing you down. When you hit a dead end, go back and try another path.
Try multiple approaches quickly without losing progress.