Support multiple agents
Run Claude Code, Codex, and future agent runtimes from the same native workspace. Pick the client and model per session without changing how you manage projects.

Supercharge your AI coding workflow with a native macOS app. Manage agent sessions, browse your project, review AI-generated diffs, and approve tool calls — all without leaving your desktop.
macOS 15+ · Apple Silicon & Intel

Live session · diff review · streaming response
Drive Claude Code sessions visually. Stream responses, approve tool calls with risk-based modals, and inspect every file the agent touches.
Run Codex inside the same interface. Switch between agents per project, with a unified session history and permission flow.
Run Claude Code, Codex, and future agent runtimes from the same native workspace. Pick the client and model per session without changing how you manage projects.

A lightweight native app with macOS materials, responsive SwiftUI views, and no bundled browser runtime. Fast to launch, fast to stream, and designed for the desktop.

Monitor active sessions, pending checks, and Claude Code or Codex usage limits from the menu bar without opening the main window.

Configure, inspect, and test Model Context Protocol servers from Settings, including local and project-scoped servers.

See every file changed in a thread with additions and deletions summarized together, even while multiple sessions keep running in parallel.

Jump directly into VS Code, Cursor, Zed, Xcode, JetBrains IDEs, Finder, Terminal, or Warp so you can manage code in the editor you choose.

Free, open source, and built for macOS. Download v1.0.1 and start driving your agents visually.