Run multiple AI coding agents in parallel — without losing control.

---

Maestro is a desktop app that turns your backlog into shipped code. Drop tasks onto a Kanban board, let agents execute them in parallel, then review and merge — all without leaving a single window.

---

How it works

Backlog  →  Ready  →  In Progress  →  Review  →  Done
              ↓             ↓              ↓
           assign        agent runs     diff viewer
           model         in isolated    hunk staging
                         worktree       one-click commit

Each task gets its own agent, its own git worktree, and its own terminal. They run in parallel. Nothing conflicts.

---

Features

Parallel agents, zero conflicts

Each task runs in an isolated git worktree. Agents don't step on each other, and you can have as many in flight as you want.

Real-time visibility

Live terminal output, a structured activity feed, and a file tree — all updating as agents work. You see exactly what's happening at every step.

Surgical diff review

When an agent finishes, you get an inline diff viewer with hunk-level staging. Accept what you want, revert what you don't, commit in one click.

Any machine, any location

Runs natively on macOS and Linux. On Windows, connect to a WSL distro or SSH into a remote Linux server — agents execute there while you work locally. Password, key, and passphrase auth all supported.

Pull work from your tracker

Sync tasks directly from GitHub Issues or Jira. Import a ticket, add instructions, hand it to an agent.

Your stack, your models

Pick the model per task. Configure MCP allowlists. Maestro stays out of the way.

---

At a glance

Capability	Detail
Kanban workflow	Backlog → Ready → In Progress → Review → Done, drag-and-drop
Parallel agents	Multiple agents run simultaneously in isolated git worktrees
Live monitoring	Real-time terminal, ACP activity feed, file tree
Diff review	Inline viewer with hunk-level staging, revert, and commit
SSH & WSL remotes	Connect to remote and WSL projects with full auth support
Issue import	Sync from GitHub Issues and Jira
Model selection	Per-task model and MCP allowlist configuration
Themes	Light, dark, and system

---

For contributors

Tech stack

Layer	Technology
Frontend	React 19, TypeScript, Vite, Tailwind CSS 4, shadcn/ui
State	Zustand + Immer, TanStack Query
Terminal	xterm.js
Desktop	Tauri 2 (Rust)
Database	SQLite (rusqlite)
SSH	russh
Protocol	ACP (Agent Client Protocol) via maestro-server sidecar
Type gen	ts-rs + tauri-specta

Prerequisites

• Rust toolchain (stable)
• Node.js 20+ + pnpm
• Tauri prerequisites for your platform
• maestro-server binary on PATH for agent execution

Getting started

pnpm install
pnpm tauri:dev

Development commands

# Frontend
pnpm dev              # Vite dev server only (localhost:5173)
pnpm build            # TypeScript check + production build
pnpm lint             # oxlint
pnpm lint:fix         # Auto-fix lint issues
pnpm format           # Check formatting (oxfmt)
pnpm format:fix       # Fix formatting

# Testing
pnpm test             # Vitest unit tests
pnpm test <pattern>   # Single test file
pnpm test:e2e         # Playwright E2E tests
pnpm test:e2e:ui      # Playwright with interactive UI

# Rust backend
cd src-tauri && cargo build
cd src-tauri && cargo test
cd src-tauri && cargo check

# Tauri
pnpm tauri:dev        # Full dev mode (Tauri + Vite)
pnpm tauri:gen        # Regenerate TypeScript bindings from Rust models
pnpm tauri build      # Production bundle (all platforms)

# Cross-compile for Windows from Linux
pnpm tauri build --debug --runner cargo-xwin --target x86_64-pc-windows-msvc

Architecture

Three Rust crates in a Cargo workspace:

• src-tauri — Tauri backend. IPC command handlers, SQLite DB, SSH tunneling, PTY management, ACP session coordination.
• maestro-server — Standalone binary (must be on PATH). ACP intermediary between Tauri and AI agents. Communicates via JSON-framed messages on stdin/stdout.
• maestro-protocol — Shared ACP protocol types.

Frontend (src/) organized as:

src/
├── components/
│   ├── common/       # App header, review modal, theme toggle
│   ├── execution/    # Agent terminal, diff viewer, worktree cards
│   ├── kanban/       # Board, columns, task cards
│   ├── project-picker/
│   ├── task/         # Task form, settings, import
│   └── ui/           # shadcn/ui primitives
├── services/         # TanStack Query service layer
├── store/            # Zustand stores
├── types/            # Generated bindings (bindings.ts) + domain types
├── utils/
│   ├── helpers/
│   └── hooks/
└── views/            # Top-level views (Kanban, Agents, Worktrees, Settings)

See CLAUDE.md for development conventions and AI agent guidelines.

---

License

Apache-2.0