A cross-platform desktop command center for orchestrating parallel AI coding agents with automated Playbooks, group chat coordination, and remote monitoring.
Maestro is an Electron-based cross-platform desktop application (macOS/Windows/Linux) positioned as an orchestration command center for AI coding agents. It serves as a pass-through orchestration layer above agents like Claude Code, OpenAI Codex, and OpenCode—without replacing agent capabilities, it solves the efficiency problem of managing multiple agents in parallel.
Core capabilities include: truly parallel AI agent execution across isolated branches via Git Worktrees; Auto Run & Playbooks for batch-executing Markdown checklist tasks from the filesystem with loop execution, full history tracking, and sessions lasting up to ~24 hours; Group Chat for multi-agent collaborative discussion with a moderator AI routing and synthesizing responses; built-in web server with QR code for mobile remote monitoring; maestro-cli providing headless mode for cron/CI/CD pipeline integration.
Developer experience features Dual-Mode sessions (AI terminal + command terminal per agent), keyboard-first design, extensible Slash Commands, regex-based output filtering, and File Explorer with @-mention file references. Analytics include Usage Dashboard (multi-dimensional stats and heatmaps), Document Graph (document knowledge graph visualization), and real-time Cost Tracking.
Built with a full-stack TypeScript codebase (99.7%), dual Vite configurations, Vitest + Playwright testing, built-in MCP Server for external integrations, SSH remote execution, and Deep Links support. Current version is v0.15.3 (Maestro Symphony), introducing community contribution mechanisms and an achievement system.
Prerequisites: at least one supported AI coding agent installed and authenticated (Claude Code / OpenAI Codex / OpenCode / Factory Droid); Git is optional.