Technology

Perplexity's New Personal Computer Brings AI to Your Own Mac Hardware

Perplexity launched Personal Computer for Mac, a new system that combines local hardware with cloud-based AI. Your Mac mini stores files securely while AI thinking happens on Perplexity's servers. You

Martin HollowayPublished 2d ago5 min readBased on 5 sources
Reading level
Perplexity's New Personal Computer Brings AI to Your Own Mac Hardware

Perplexity's New Personal Computer Brings AI to Your Own Mac Hardware

On April 17, 2026, Perplexity launched Personal Computer for Mac. It's a new system that turns a Mac mini into a persistent AI assistant — one that runs 24/7, stays connected to your files and apps, and combines computing power from both your local hardware and the cloud.

Think of it as a digital proxy living inside Perplexity's desktop app. You can ask it to handle tasks, organize files, and automate workflows from any of your devices, while it maintains access to everything stored on that dedicated Mac. Windows support is coming, with a waitlist already open.

How It Works: Local and Cloud Together

Personal Computer uses an interesting split design. Your files and sensitive data stay on the Mac mini in your home or office. The actual AI thinking happens on Perplexity's servers in the cloud. They're connected securely, so you get the speed and smarts of powerful AI models without sending everything to the internet.

The system defaults to Claude Opus 4.7 as its main AI brain, but you can also choose GPT-5.4 or Claude Sonnet 4.6, depending on what you're asking it to do. Some tasks need raw power; others need precision. This flexibility lets different jobs use the model that suits them best.

Personal Computer also learns from how you work. It builds what Perplexity calls "persistent memories" — a record of your habits, preferences, and recurring tasks. Over time, it needs less manual instruction to handle repetitive work.

Safety and Control

Before the system does anything risky — like deleting files or changing important settings — it asks for your permission. You stay in the loop. This matters, because as AI systems become more autonomous, humans still need the final say on consequential actions.

The local network design also means the AI system doesn't have to talk to the internet for every single operation. It operates within your existing network security setup, which may appeal to people who want tight control over what data leaves their office or home.

Enterprise Version Too

Alongside Personal Computer, Perplexity rolled out Computer for Enterprise, aimed at companies rather than individuals. It includes built-in compliance features, can route tasks to different AI models based on what's safest or most cost-effective, and plugs into the software tools companies already use.

What This Means: Trade-Offs

The broader context here is worth considering. During the early days of cloud computing — around 2008 to 2010 — many organizations didn't want to commit entirely to the cloud. They had sensitive data they wanted to keep on premises, but they also wanted the computing power that cloud providers offered. So they built hybrid systems: data at home, processing in the cloud. It was more complex than either pure approach, but it gave them flexibility during a transition.

Personal Computer is solving a similar problem for the AI era. Users want the reasoning power of advanced language models, but they don't want to hand over control of their files and their computing environment. This system lets them have both, at the cost of some extra complexity — you need a dedicated Mac mini always plugged in and connected to the network.

In this author's view, the real test will be whether people find that trade-off worth it. You're buying hardware, using electricity to keep it running, and accepting a more complicated setup. In return, you get data that stays local and an AI assistant that's always available. Some people will see that as essential infrastructure. Others will prefer simplicity over control.

The broader trend in AI right now is everyone experimenting with different points in this same spectrum — different balances between capability, ease of use, and control. Personal Computer is one approach. It prioritizes your data staying yours, while still letting you access powerful AI models. Whether it catches on will depend partly on how much that matters to the people it's built for.