Maestro’s offline mode wasn’t on our roadmap. It started with a phone call from a friend who does disaster relief logistics. He needed software tools in places where the internet is the first thing to go and the last thing to come back.
He was describing exactly the kind of work Maestro was built for — well-specified, practical software with clear requirements. The only problem: every AI development tool on the market required a persistent internet connection.
What airplane mode does
When you enable offline mode, Maestro swaps two dependencies:
- Cloud AI models become local Ollama models — running entirely on your machine
- GitHub becomes a local Gitea server — spun up automatically in Docker
Everything else stays the same. The PM still interviews you. The Architect still plans and reviews. The Coders still write code and submit PRs. The workflow is identical; only the infrastructure changes.
Who actually uses it
We built it for disaster relief. But the use cases that emerged surprised us:
- Government and defense contractors who can’t send code to external APIs
- Teams in regions where frontier model APIs are expensive or unavailable
- Developers who just want to work on a plane without worrying about connectivity
- Privacy-conscious users who want zero data leaving their machine
The tradeoffs
Let’s be honest: local models aren’t as capable as frontier cloud models. You’ll see more PR rejections, more iterations, and longer build times. A project that takes two hours with Claude and ChatGPT might take six hours with local models.
But six hours of fully offline, fully private development — with a complete PR history and test coverage — is still remarkable. And for the use cases where offline is a requirement, not a preference, there’s simply no alternative.
Syncing back
When you’re back online, maestro --sync pushes everything to GitHub. Your local Gitea history becomes real GitHub PRs and merges. Nothing is lost in the transition.