Give ChatGPT
or Claude
control of
your iPhone.

iPhonePilot provides packaged local control tools, an installation checklist your AI agent can follow, profession-specific workflow prompts, and support. The Skill brings the required installation and configuration steps together, rather than leaving you to assemble the setup instructions yourself.

Your AI can use screen viewing, taps, text entry, and scrolling to navigate supported interfaces. The workflow examples illustrate how those actions can support testing, content preparation, and operations. Completion depends on the app, starting state, and task complexity. The guarantee covers basic device control.

Supervision depends on the task and the app. Observe the first run and define stopping points in your instructions. A request to stop before publishing or purchasing is an instruction to the AI, not a verified technical approval lock. Fully unattended execution and a fixed number of hours saved are not promised.

If ChatGPT or Claude cannot see the screen and perform basic taps, typing, and scrolling on your supported setup, request a full refund of the iPhonePilot AI purchase price within 14 days of purchase. LLM costs and arbitrary scenario outcomes are not covered.

Support is included in the $79 purchase. Response times and future update coverage are not currently specified. The one-time price covers the purchased product; it does not establish access to every future version.

The Skill includes the required installation and configuration instructions. Provide them to Codex or Claude Code on your computer and ask the agent to follow the checklist. The agent shows its installation plan and performs the approved steps. You connect the iPhone and handle any permissions, authentication, device prompts, or fallback steps that require your involvement. A standard web chat alone cannot install software on your computer.

Unlock the phone, connect it with a USB data cable, and confirm Trust This Computer when prompted. Enable Developer Mode if your setup requires it; this may involve a restart. Sign in to the apps you want to use yourself. Enter passcodes and passwords directly on the device, rather than giving them to the AI. Keep the phone unlocked while the agent is working.

Start with a small test: ask the agent to detect the connected iPhone, capture its screen, open Safari, capture the screen again, and stop. Then check taps, text entry, and scrolling in a non-sensitive task before using a longer workflow. The included guide walks through the initial checks and common connection issues.

Founder tests reached first connection and basic AI control in 10–20 minutes on tested configurations. Your environment may require additional setup. Older iOS versions may need an additional device-control component and extra on-device steps. The guide explains the applicable setup path. This is the time to establish control, rather than a completion time for a full workflow.

Founder testing includes iPhones starting with iPhone 8, connected by USB to Windows or macOS computers using Codex or Claude Code. Confirm your specific OS version and installation requirements with support; the device model alone does not establish compatibility.

The founder-tested route uses a USB-connected iPhone with Codex or Claude Code on your computer. iPhonePilot makes local screen-viewing and action tools available to the agent. The model interprets screen information and directs actions on the phone. A standard ChatGPT or Claude web chat or ZIP upload alone does not establish this connection. LLM usage is paid separately.

There is no additional per-device license fee for supported iPhones. This does not establish unlimited simultaneous operation, team seats, or resale rights.

You define the task, success criteria, and actions the AI may take. Review its observations before making release or business decisions. Specific interruption controls and enforced approval mechanisms have not yet been documented for this release; do not assume that every consequential action is technically gated.

Your chosen AI integration receives screen information so the model can interpret the interface and direct actions. Local phone-control tools do not mean that model processing stays on your computer. Data handling depends on the integration and model provider; account and screen information should be appropriate for that environment.

Email uroborisstore@gmail.com for support or to request a refund. If basic screen access, taps, typing, or scrolling do not work on your supported setup, make the request within 14 days of purchase. The refund covers the iPhonePilot purchase price; LLM costs and individual workflow outcomes are excluded.