# ace

> ace is a free, open-source app where Claude Code, Codex and Cursor work as one team across your Mac, Windows and Linux machines, with phone apps.

ace is a free, open-source app that runs your coding agents as one team on Mac, Windows and Linux, with ace for iPhone and iPad to follow and steer them (Android is coming soon). It drives the agent CLIs you already have logged in (Claude Code, Codex, Cursor, OpenCode, Antigravity, Pi and any ACP agent), so many tasks run at once, each in its own thread and, if you like, its own git worktree with a built-in diff review. Any agent can hand work to agents on other providers and on your other machines, and a thread shows as done only when every agent under it has finished. Agents get one MCP toolset in every CLI: they delegate to and wait on other agents, use a real browser, operate desktop apps through computer use, and test on iOS Simulators and Android emulators, including ones on another of your machines. Sign in to several accounts per provider and move a thread to another when one hits its limit. Sign in to an ace account and every machine on it opens from your phone, browser or other computers, from anywhere, through an end-to-end encrypted relay.

## Key facts

- ace is a free, open-source (Apache-2.0) desktop app for Mac, Windows and Linux, with apps for iPhone and iPad.
- ace runs the coding-agent CLIs you already use: Claude Code, Codex, Cursor, OpenCode, Antigravity, Pi and any Agent Client Protocol (ACP) agent.
- ace never asks for, stores or proxies provider credentials; each agent signs in through its own CLI, on your existing subscription.
- In ace, any agent can delegate work to agents on other providers, for example Claude Code handing a task to Codex, and wait for the results.
- Agents in ace can start threads and use browsers, iOS Simulators and Android emulators on your other linked machines.
- A thread in ace is done only when every agent and background command under it has finished; it shows when any of them needs you.
- ace supports several accounts per provider and moves a thread to another account when one reaches its usage limit.
- Every agent in ace gets the same MCP tools: delegation, a real browser, computer use, simulators and emulators, and automations.
- Each ace thread can work in its own git worktree, with a diff review beside the conversation.

## Links

- [Download for Mac](https://github.com/arpan404/ace/releases/download/v0.0.19/ace-0.0.19-arm64.dmg): ace 0.0.19 for macOS on Apple silicon, 239 MB
- [Download for Windows](https://github.com/arpan404/ace/releases/download/v0.0.19/ace-0.0.19-x64-setup.exe): ace 0.0.19 for Windows on x64, 208 MB
- [Download for Linux](https://github.com/arpan404/ace/releases/tag/v0.0.19): the latest release page, for Linux on x64 (AppImage)
- ace for iPhone and iPad (Android coming soon): follows and steers agents on every machine you pair or sign in to your ace account
- [Release notes](https://github.com/arpan404/ace/releases/tag/v0.0.19)
- [Source code on GitHub](https://github.com/arpan404/ace): Apache-2.0
- [All releases](https://github.com/arpan404/ace/releases)
- [Everything on this site in one file](https://acehq.dev/llms-full.txt)

## Pages

- [Home](https://acehq.dev/index.md): What ace is and what it does. Web page: https://acehq.dev/
- [Download](https://acehq.dev/download.md): ace for Mac, Windows and Linux, and ace for iPhone, iPad and Android. Web page: https://acehq.dev/download/
- [Features](https://acehq.dev/features.md): Everything ace adds to the coding agents you already use. Web page: https://acehq.dev/features/
- [Agent orchestration](https://acehq.dev/features/orchestration.md): Let one coding agent hand work to others: Claude Code delegates to Codex, Cursor or any agent, on any of your machines, and waits until the whole team is done. Web page: https://acehq.dev/features/orchestration/
- [Computer use](https://acehq.dev/features/computer-use.md): Let Claude Code, Codex and other agents use your desktop apps in the background, with a live view, Take over and Stop, and apps you keep off limits. Web page: https://acehq.dev/features/computer-use/
- [Fleet](https://acehq.dev/features/fleet.md): Link your machines so agents share each other's threads, browsers, simulators and emulators, move work between them, and follow every thread from iPhone or iPad. Web page: https://acehq.dev/features/fleet/
- [iPhone and iPad](https://acehq.dev/features/mobile.md): Follow and steer every agent from iPhone or iPad: approvals, diffs, terminals, the live browser, your desktop's simulators and Live Activities on the Lock Screen. Web page: https://acehq.dev/features/mobile/
- [Remote access](https://acehq.dev/features/remote-access.md): Sign in once and every machine on your ace account opens from your phone, browser or laptop, anywhere, through an end-to-end encrypted relay. Web page: https://acehq.dev/features/remote-access/
- [Accounts and limits](https://acehq.dev/features/accounts.md): Sign in to several accounts per provider, pick one per thread, see usage and limits, and move a thread to another account when one runs out. Web page: https://acehq.dev/features/accounts/
- [Browser use](https://acehq.dev/features/browser.md): Agents open your dev server in ace's browser, click through it and read the console and network log, while you watch the same tab and can take over. Web page: https://acehq.dev/features/browser/
- [Devices](https://acehq.dev/features/devices.md): Run iOS Simulators and Android emulators inside ace. Use them yourself or hand one to an agent to test your app, with live screen, UI tree and logs. Web page: https://acehq.dev/features/devices/
- [Worktrees and review](https://acehq.dev/features/review.md): Give every agent thread its own git worktree, review its diff beside the conversation, send line comments back, then commit, push and open the pull request. Web page: https://acehq.dev/features/review/
- [Thread status](https://acehq.dev/features/status.md): ace shows whether each agent thread is working, waiting, failed, done or needs you, counting every subagent and background command it started. Web page: https://acehq.dev/features/status/
- [Agents](https://acehq.dev/agents.md): The coding agents ace runs: built in, and any ACP agent. Web page: https://acehq.dev/agents/
- [Claude Code](https://acehq.dev/agents/claude-code.md): Run Claude Code in ace: Anthropic's coding agent, run from its own command-line tool. Web page: https://acehq.dev/agents/claude-code/
- [Codex](https://acehq.dev/agents/codex.md): Run Codex in ace: OpenAI's coding agent, run from its own command-line tool. Web page: https://acehq.dev/agents/codex/
- [Cursor](https://acehq.dev/agents/cursor.md): Run Cursor in ace: Cursor's coding agent, run from its command-line tool. Web page: https://acehq.dev/agents/cursor/
- [OpenCode](https://acehq.dev/agents/opencode.md): Run OpenCode in ace: The open-source coding agent that works with many model providers. Web page: https://acehq.dev/agents/opencode/
- [Antigravity](https://acehq.dev/agents/antigravity.md): Run Antigravity in ace: Google's Antigravity coding agent. Web page: https://acehq.dev/agents/antigravity/
- [Pi](https://acehq.dev/agents/pi.md): Run Pi in ace: The Pi coding agent, run from its own command-line tool. Web page: https://acehq.dev/agents/pi/
- [MCP tools](https://acehq.dev/mcp.md): The MCP tools every agent gets in ace, in every CLI. Web page: https://acehq.dev/mcp/
- [Security and privacy](https://acehq.dev/security.md): Where your code goes, what ace holds, and what it never touches. Web page: https://acehq.dev/security/
- [FAQ](https://acehq.dev/faq.md): Questions people ask before trying ace. Web page: https://acehq.dev/faq/
- [Privacy policy](https://acehq.dev/privacy-policy.md): What ace's account, relay and website collect, how it's used and how to delete it. Web page: https://acehq.dev/privacy-policy/
- [Terms of service](https://acehq.dev/terms-of-service.md): The terms for using ace's account, hosted relay and other services. Web page: https://acehq.dev/terms-of-service/

## Supported agents

- Built in: Claude Code, Codex, Cursor, OpenCode, Antigravity, Pi.
- Through the Agent Client Protocol (ACP): Amp, Cline, Codebuddy Code, Cortex Code, DeepAgents, Devin, Gemini CLI, GitHub Copilot, goose, Grok Build, Junie, Kilo, Kimi CLI, MiniMax Code, Mistral Vibe, Qoder CLI, Qwen Code, and any other ACP agent.
- ace uses each agent's own logged-in CLI. It never asks for, stores or proxies provider credentials.

## Thread status

A thread's status covers every subagent and background command it started.

- Working: The agent, or one of its subagents, is running.
- Approve: It is stopped on a permission or a question for you.
- Waiting: Its turn ended, but subagents or commands still run.
- Failed: It stopped on an error you should read.
- Done: Every agent and command in the thread has finished.