Maestri

Meet Maestri. An infinite canvas where your coding agents work in concert.

Developers love terminals. That doesn't need to change. Everything around them does.

AI agents have changed the game. They write, refactor, and test code faster than any human ever could. But here's the catch: the bottleneck has shifted from the code to the coder.

You can spin up three agents at once, but you're still alt-tabbing through a mess of terminal windows, losing context, forgetting which agent is doing what. The tools caught up. Your workspace didn't.

Everything you need, no distractions.

• Agents that talk to each other: Connect terminals together and let agents prompt each other directly. One agent can delegate tasks, ask questions, or hand off work without you lifting a finger.
• Draw your thinking: Sketch architecture diagrams, draw arrows between components, map out flows, right on the canvas, next to the agents doing the work. Hand-drawn shapes that feel natural and fast.
• What if your terminal had no edges? Pan freely. Zoom in on the agent that needs attention. Zoom out to see the full picture. Group related work spatially, the way your brain already thinks about projects.
• Spawn agents anywhere: Drop a new terminal wherever you need it. Each one runs its own agent: Claude Code, Codex, OpenCode, or just a plain shell. See all of them at once. No tabs. No alt-tab. Just space.
• Notes, written by your agents: Connect a terminal to a sticky note and let your agent write directly to it. Progress updates, summaries, decisions, captured automatically while you focus on what matters.
• One canvas per context: Separate canvases for separate projects. Jump between terminals and workspaces with tmux-style shortcuts that feel instant. Each workspace remembers exactly where you left off.
• Jump straight to the code: One click to open the project in VS Code, Zed, Xcode, or your favorite editor. Go from watching an agent work to editing the code yourself, instantly.

Always on your shoulder. Never in your way.