What is apple-ads-cli?

Apple Ads CLI by Lily is a command line tool for planning and managing Apple Search Ads campaigns, built with AI assistants like Codex and Claude Code in mind. It is an independent, unofficial project and is not affiliated with or endorsed by Apple. Instead of exposing raw API calls, it groups the steps of setting up a campaign, such as creating the campaign itself, an ad group, keywords, negative keywords, a cost per acquisition goal, and a creative attachment, into a single business focused command. By default the tool returns structured JSON output and treats any command that would write changes to Apple Ads as a dry run, so it shows what would happen and asks for confirmation before actually making changes. This is meant to let an AI agent explain the plan in plain business terms, such as the suggested budget, bid, and keywords, and let a human approve, adjust, or cancel before anything is sent live. The tool works against Apple's current Campaign Management API version 5, while also preparing support for Apple's newer Platform API, and it hides the differences between the two behind a single interface so a user does not need to learn both. You can install it with Homebrew, a hosted install script, by building from source with Go, or by downloading a release binary. Configuration happens through environment variables or a local config file holding your Apple Ads credentials, and your private keys stay on your own machine. The project also ships ready made skill files for Codex and Claude Code that teach those agents how to use the tool safely, always previewing a plan before writing. On top of the core Apple Ads features, which are free and open source, you can optionally sign in to a hosted service called Lily to unlock keyword level revenue analytics, pulling in data from RevenueCat or AppsFlyer to show return on ad spend. Without that optional login, the tool still works for planning and managing campaigns, it simply cannot calculate revenue based metrics. The project is written in Go and released under the MIT license.