Perplexity Brings Its Autonomous AI Computer To Android Users

The boundary between human intent and software execution just shifted again. Following its debut on iOS earlier this month, Perplexity has officially brought its "Perplexity Computer" to Android, turning the world’s most popular mobile operating system into a launchpad for autonomous AI agents.

From Chatting to Doing

Perplexity Computer isn't a traditional chatbot; it’s an orchestration layer that functions like a high-end digital intern. When a user requests a complex task—such as researching competitors, synthesizing that data into a slide deck, and emailing it to a team—the system doesn't just provide links. It breaks the project into subtasks, delegates them to a curated suite of 19 specialized models, and executes the entire workflow inside an isolated, secure Linux sandbox. By leveraging powerhouses like Claude Opus 4.6 for reasoning and Gemini for research, the system manages the heavy lifting in the cloud, keeping your actual device battery and processor clear of the load.

The genius of the system lies in its persistence. Unlike a standard query that ends when you close the tab, these AI agents operate asynchronously. You can start a task on your desktop, leave for a meeting, and check the status—or redirect the workflow—from your Android phone. CEO Aravind Srinivas has been clear that this is a fundamental shift in computing: we are moving from the era of 'chatting with AI' to the era of 'hiring AI to get things done.'

The Future of Asynchronous Work

By making this technology available on mobile, Perplexity is betting that the most valuable AI assistant is the one that works while you aren't looking. This is the transition from 'read-only' mobile interfaces to mobile command centers for multi-hour, multi-app processes. While the $200-a-month price tag for a 'Perplexity Max' subscription currently limits the feature to power users and enterprises, the infrastructure being built here is a blueprint for how we will all interact with digital software in the coming years.

The real challenge, and the true test for Perplexity, will be maintaining trust as these agents gain more autonomy. As they begin interacting with Salesforce, Slack, and email accounts on our behalf, the margin for error shrinks. Perplexity is countering this with a 'user-in-the-loop' approach, requiring approval for sensitive actions, which bridges the gap between raw AI speed and human safety. We are witnessing the birth of the digital employee, and it just gained the ability to follow you anywhere your phone goes.