Perplexity Launches 'Computer' AI Agent
AI startup Perplexity has launched "Perplexity Computer," a multi-model system that can autonomously create and execute complex workflows. The platform represents a move towards more capable AI agents that can perform entire sequences of tasks, not just answer single queries, potentially reshaping knowledge work.
Perplexity's new "Computer" agent is a significant step beyond single-function AI chatbots. It operates as an orchestrator, breaking down a user's goal into smaller pieces and assigning each subtask to the most suitable AI model from a pool of 19 specialized systems. This allows it to handle complex, multi-step projects from a single prompt, such as building a competitive intelligence dashboard or deploying a new landing page. The system is designed to "use software the way a human would," according to CEO Aravind Srinivas. Instead of just providing answers, Perplexity Computer interacts with applications, manipulates files, and pulls data from APIs to deliver a finished product without needing manual intervention between steps. For instance, it uses vision models to interpret what's on a screen and language models to reason through the necessary actions to complete a task. This multi-model approach is a key differentiator. While many platforms are built around a single primary AI, Perplexity Computer dynamically routes tasks to different models based on their strengths—using models like Gemini for research, Claude for coding, and Grok for quick lookups. This strategy acknowledges that different models excel at different types of work. The agent is available to Perplexity's "Max" subscribers for $200 per month, a price point that puts advanced agent capabilities in the hands of individuals and power users, not just enterprise clients. Unlike open-source agents like OpenClaw that run on a user's local machine, Perplexity Computer is entirely cloud-based, offering centralized safety and automatic updates. Perplexity AI was founded in August 2022 by Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski, with backgrounds at OpenAI, Google, and DeepMind. The company has seen rapid growth, reaching a valuation of over $1 billion by April 2024 and securing investments from notable figures like Jeff Bezos and Nvidia.
