OpenClaw launches a mobile app for controlling AI agents on Android and iOS

The OpenClaw AI agent now has an official mobile app for Android and iOS. Users can control the autonomous agent from their smartphone and grant it access to the device's camera, screen, location, photos, calendars, and reminders.

What the mobile app can do

The new app turns the smartphone from a simple remote control into a full sensory channel for the agent.

• Runs on Android and iOS
• Gives the agent access to the device camera
• Lets the agent work with on-screen content
• Shares the user's location data
• Opens up access to photos
• Syncs with calendars and reminders

This set of permissions goes far beyond a typical chatbot. Instead of just answering text prompts in a chat window, the agent can see the user's surroundings through the camera, orient itself using location data, and manage the user's personal schedule and reminders on its own. Until now, this level of device integration was largely limited to voice assistants like Google Assistant or Siri, which had access to some of the same data but did not act autonomously on the user's behalf.

Why autonomous agents need device access

Over the past couple of years, AI agents that act rather than just respond with text — opening apps, clicking through interfaces, reading data from a phone's sensors — have gained momentum alongside growing interest in browser and "computer" agents. OpenAI demoed its Operator browser agent, Anthropic is building out "computer use" features for its Claude models, and Google is experimenting with its Project Mariner agent. Across the industry, the focus is shifting from chatbots to systems that act on their own.

OpenClaw's mobile app follows the same path: instead of a separate browser extension or desktop client, the agent gets direct access to the device users carry in their pocket for most of the day — the one that holds their most personal data, including photos, messages, location, and schedule.

How this changes the way the agent is used

For users, this means the agent could potentially handle end to end the tasks that used to require manually switching between apps and screens: photograph a document with the camera, check the calendar, find the right photo in the gallery, and set a reminder — all from a single request. The breadth of the requested permissions is what makes the release notable. OpenClaw is one of the few mass-market mobile AI agents to officially request the full set of sensitive device permissions in a single app, rather than piecemeal through separate integrations.

The launch of an official client also signals the project's maturity. Instead of working with a third-party API or a custom build for a specific task, users get a ready-made app. That typically lowers the barrier to entry for less technical users, while also raising the bar for how the developer protects the data passed to the agent.

What it means

Extending autonomous agents' permissions to a smartphone's camera, location, and personal data speeds up the shift from chatbots to full-fledged digital assistants that can act for users on their own devices. It also raises security concerns: the more sensors and personal data an agent can reach, the higher the cost of a mistake, a leak, or a breach.