Skip to content

Summary

An open-source Mac app that gives Claude Code a real browser, a live iOS Simulator and its own git worktree per chat.

Screenshots

Description

A terminal is a thin place for a coding agent to live. Superagent is a native macOS app that gives one a workspace instead: a real browser it can drive, an iOS Simulator it can tap and swipe, a git worktree per conversation, and chats that survive quitting the app.

The browser pane is a genuine browser session with your existing logins, so the agent can sign in to a staging environment or a dashboard the way you would rather than fighting a headless shell. Alongside it, a real iOS Simulator streams from its own framebuffer — nothing to install — so an agent working on a mobile build can tap through the result. A dual-viewport mode renders the same page at desktop and phone widths in one live session, which makes "does the hero hold up on mobile?" a question the agent can answer itself.

Each chat runs on its own branch through a git worktree, so parallel experiments never touch your main checkout. Sessions and their context persist across restarts, and a companion iPhone app puts you in the loop remotely: approve a step or redirect a run from your phone, over an end-to-end encrypted relay.

Superagent supports Claude Code today, with Codex and Antigravity listed as coming next, and it drives whatever agent subscription you already pay for rather than reselling tokens.

Requirements: macOS on Apple Silicon. Free, open source and MIT licensed, distributed as a signed DMG from GitHub Releases.

Reviews

Similar App Suggestions

App: seendiff

Abhishek Iyer

New

A local, keyboard-driven diff viewer built for the huge changes coding agents produce - and it remembers what you have already seen.

Coding & DevelopmentFree
New

Write an agent as a TypeScript function and deploy it in seconds onto KVM-isolated Linux microVMs, with the loop, sessions, streaming and versioning managed for you.

Coding & DevelopmentPaid

Open-source inference engine that runs LLMs, vision and speech models fully on-device — no API keys, no cloud calls, from Swift to Godot.

Coding & DevelopmentFree

App: Lenz

Lenz IO

New

Audit-grade fact-checking API for AI products — eight models across five stages, adversarial debate, and a full citation trail behind every verdict.

Coding & DevelopmentFreemium

Browser panel with 70 diagnostics that names a page's stack, extracts its design tokens and scores how much of it reads as AI-generated.

Coding & DevelopmentFreemium