ai-maestro

Why we included this project

Anyone who has run several terminal-based coding agents at once knows the pain of being the human message bus, copying context from one session and pasting it into another. AI Maestro is built to remove that job: a single dashboard shows every agent on every machine, and it works with Claude Code, Codex, Aider, Cursor, and other terminal agents rather than locking you into one vendor. The Agent Messaging Protocol is the interesting part, letting agents send each other signed, prioritized messages so a research agent can hand findings directly to a writing agent without you relaying them. Persistent memory, code-graph queries, and team coordination features like war rooms and a Kanban board round it out, which is what makes it work as a control plane for a small AI workforce. If you are past the point of juggling terminals and want agents to coordinate like an actual team, this is worth a look.