Perplexity launches Teammates AI agents on Mac for Max subscribers

From answers to actions

Perplexity has officially launched Teammates within its Mac application. This new feature is specifically available to users who subscribe to Perplexity Max, which is the company's highest paid tier. By bringing these tools to the desktop, the company aims to place AI capabilities where most professional work actually happens.

The release represents a significant shift in how the platform functions for its users. While the service began as an answer engine, it is now moving toward agentic products. These are systems designed to run work in the background rather than requiring a user to interact with every single step of a process.

By moving these tools out of the browser and into a dedicated Mac app, Perplexity is targeting a more integrated workflow. This transition suggests that the company wants users to view the platform as a workspace for completing projects. It allows for a more seamless experience for people who spend their day moving between different applications and tasks.

How Teammates work

Teammates are defined as AI agents that are built to own a specific task. Unlike a standard chatbot that provides a single response to a single prompt, these agents are designed to carry out a job from start to finish. They can operate independently to handle smaller assignments or work in tandem with other agents.

When multiple Teammates work together, they can split a larger job among themselves to increase efficiency. This collaborative structure allows for more complex workflows to be managed by the AI. The system is designed to coordinate ongoing work, allowing the agents to maintain context over a series of actions.

Perplexity has identified several specific areas where these agents can be applied. They highlight quality assurance, customer service, research, and growth marketing as primary examples. However, the company noted that the agents are capable of handling many other types of uses beyond these specific categories.

The strategy of exclusivity

The decision to limit the initial release of Teammates to Max subscribers follows a common pattern in the AI industry. High-end features often debut with the most expensive tiers first because the computing costs for running agents are significantly higher. This allows companies to manage their overhead while testing the technology with their most dedicated users.

By starting with the Max tier, Perplexity can monitor how these agents perform in real-world scenarios. This feedback loop is essential for refining the reliability of autonomous systems before they are scaled. It ensures that the most complex features are polished before they reach a broader audience.

Currently, there is no word on when these features might become available to Pro or free users. The company has also not confirmed if the Mac app functionality will expand to other operating systems. For now, the focus remains on providing a high-performance experience for the top-tier subscriber base.

Competing in the agent space

The introduction of Teammates puts Perplexity in direct competition with the growing field of multi-agent workspaces. These are environments where different AI models are organized to work together on corporate goals. By offering this, Perplexity is attempting to capture a larger share of the professional productivity market.

For teams already paying for the Max subscription, this update changes the value proposition of the service. It transforms the platform from a place where users go to look things up into a tool where they can hand off recurring work. This shift targets the way businesses manage repetitive and time-consuming administrative or creative tasks.

The move signals a broader trend in the industry toward 'agentic' workflows. Instead of humans directing every click, the human provides the goal and the AI manages the execution. This could fundamentally change how small teams and large corporations interact with their software stacks on a daily basis.

What comes next for agents

As Perplexity continues to develop Teammates, the focus will likely remain on expanding the variety of tasks these agents can perform. The goal is to make the agents as versatile as possible so they can handle diverse needs across different industries. This evolution will depend on how well the agents can coordinate and handle complex, multi-step instructions.

The company's move to the Mac app suggests that desktop integration is a priority for the next phase of growth. Users who work on laptops and desktops will have a more native experience than those relying solely on web interfaces. This could lead to deeper integration with other files and applications on the local machine.

While the current rollout is limited, the trajectory points toward a more agent-centric future for the platform. Whether these tools eventually reach every user remains to be seen, but the foundation for a multi-agent workspace is now firmly in place for Max subscribers.

Technical overview

To understand how these agents function, it helps to distinguish between a standard LLM (Large Language Model) and an agent. A standard model predicts the next word in a sequence to answer a prompt. An agent uses that same underlying model but adds a layer of reasoning to plan and execute a series of actions to reach a goal.

In the case of Teammates, the 'agentic' nature means the AI can break down a request like 'research this company' into smaller sub-tasks. It might first search for the company's website, then look for recent news, and finally summarize the findings. Each of these steps is a separate action that the agent performs autonomously.

The ability for multiple agents to collaborate is a key distinction in this release. It allows for a division of labor where one agent might focus on data collection while another focuses on synthesis. This multi-agent approach is designed to handle the complexity of professional work that a single prompt cannot easily describe.