Muster: The native cockpit for AI coding agents

Hey PH — maker here. Muster started as my own frustration: babysitting AI coding agents meant juggling branches so they wouldn't collide, worrying about what an agent might touch in my databases, and copy-pasting context between a terminal, a DB client, and an issue tracker all day. The agents were fast; everything around them was chaos. So I built the workspace I wanted: every task runs in its own git worktree, agents only reach data through MCP with permissions I set explicitly, and nothing ships until I've reviewed the diff hunk by hunk. SSH, a five-engine DB studio, and tracker imports live in the same native app, so the loop from "ticket" to "merged PR" happens in one window. Two things people tend to ask about: • Offline/privacy: code, databases and credentials never leave your machine. License checks verify offline — no account, no telemetry on your work. • Pricing: $39 one-time per device. No subscription, no seats math, no "AI credits". You buy it, you own it. If you try it, the fastest way to feel it: import an issue, let an agent work it in a worktree, then review the diff and open the PR without leaving the app. I'd love to hear what your agent workflow looks like today — and what's the one thing you would never let an agent touch. I'll be here all day answering questions.