Docs
Learn Mstro in an afternoon.
Short guides on how Mstro works — install, connect a project, tour the Security Bouncer, and run parallel AI agents on git worktrees.
Start here
Core features
The Security Bouncer
How the 2-layer tool-approval system (pattern matching + Claude Haiku) lets long-running AI agents proceed safely without constant permission prompts.
Read →The PM Board
Turn one prompt into a kanban of tasks and let AI agents work them in parallel on separate git worktrees.
Read →Git worktrees in Mstro
How Mstro uses git worktrees to run many AI agents on the same repo at the same time without collisions.
Read →
Can’t find something?
Check the FAQ, the comparison pages, or email bravo@mstro.app.