AI Agents Need a Body. Cloud Phones Might Be It.

Everyone's talking about AI Agents. But there's a question most people skip over: Where does the Agent actually run? Text generation, summarization, reasoning — those happen in the model. But the moment you ask an Agent to do something in the real world — open an app, scroll a feed, tap a button, fill a form — it needs an environment to act in. That environment is the missing piece most Agent discussions ignore. The problem with browser-only automation Most Agent frameworks today operate inside browsers or via APIs. That works for a lot of tasks. But huge portions of real-world workflows live inside mobile apps — and those apps don't have APIs you can just call. Instagram, TikTok, WhatsApp, Shopee, Lazada — the interfaces billions of people use every day are mobile-first, and largely closed to traditional automation. Enter the cloud phone A cloud phone is an Android device running on a remote server. You access it through a browser. The apps, storage, and processing all live in the clo