TL;DR: Perplexity Computer local on Mac is now an option: Perplexity Computer can run fully on device. Perplexity announced on October 20, 2026 that its Computer agent can run the AI model, the agent harness and task execution entirely on a Mac with Apple silicon, working across local files and apps like Notes, Mail, Messages and Numbers without sending that data to the cloud. It runs Qwen 3.6 or PPLX Pro, a post-trained version of Qwen. Perplexity recommends 36GB of unified memory for fully local mode, and the feature is for Pro, Max and Enterprise subscribers.
Most AI agents do their thinking on someone else’s servers. Perplexity is betting that a growing share of work, especially work involving private files, should stay on the user’s own machine. With Perplexity Computer local on Mac, the company says the whole task, from planning to finished output, can happen on device, with the cloud used only when the user allows it. Here is what the update does, what you need to run it and what to keep in mind.
What Perplexity announced
In a blog post, Perplexity said it is “launching Computer fully local on Mac, making it possible to run the AI model, harness, and task execution entirely on device.” When given permission, Computer can work across local files and Mac apps “without sending personal data to the cloud,” which the company says gives users more control over sensitive data and credit spend.
This is the third step in a sequence. Perplexity says it first brought Computer to local files and native Mac apps with Personal Computer for Mac in May, then launched Hybrid Compute last month, which combines cloud reasoning with local models that handle private information. The new mode goes further by keeping the entire task on the Mac.
Fully local vs Hybrid Compute
The Perplexity Mac app now offers two ways to run Computer. Perplexity’s product page lays out the differences:
| Fully local | Hybrid Compute | |
|---|---|---|
| Where the task starts | On the Mac | In the cloud |
| Who does the reasoning | A local model (Qwen 3.6 or PPLX Pro) plus the local agent | Cloud models for research and reasoning; a local model for private files |
| Private files | Processed on device, not sent to cloud models | Handled locally; a “Privacy Gate” can keep, mask or ask before sharing personal information |
| Recommended memory | 36GB unified memory | 32GB unified memory |
| Best for | Sensitive work that does not need live web data | Tasks mixing private files with current research |
Perplexity gives an example of each: a portfolio analysis built from private holdings and downloaded market data could run entirely on device, while a deal review combining current market research with confidential financials could use Hybrid Compute.
What it can do on your Mac
Perplexity says that with permission, local Computer works across files and supported apps including Notes, Messages, Mail and Numbers. The company’s own examples include:
- Reading a project to-do list in Notes, checking a teammate’s budget request in Messages and pulling spending figures from a spreadsheet to prepare a status update.
- Reviewing private portfolio holdings against investment guidelines and updating an Excel workbook and Keynote deck.
- Turning confidential discovery notes and an internal rate sheet into a client proposal.
According to 9to5Mac, Perplexity also gives the example of reviewing recent messages and emails, identifying deadlines and creating an action list in Notes. 9to5Mac reports the mode can run tasks without an internet connection.
Escalating to the cloud, only when allowed
Local does not mean cut off. Perplexity says users can authorize Computer to escalate to the cloud for current information, browser use and connected apps, and that the local orchestrator can call on one of 15+ frontier models for advanced reasoning. Perplexity’s product page also clarifies a subtle point: “fully local” describes where the AI runs. Sending a message or meeting invitation still uses that app’s own service.
On security, Perplexity says code and tool execution run in isolated sandbox environments with controlled access to files and connected apps, and that the same protocols apply to the cloud version. The company has previously open-sourced its agent sandbox technology.
Requirements and how to set it up
What you need
- A supported Mac with Apple silicon.
- 36GB of unified memory recommended for fully local mode (32GB for Hybrid).
- A Perplexity Pro, Max or Enterprise subscription, per 9to5Mac.
- The latest Perplexity Mac app.
Setup steps
- Download and open the latest Perplexity Mac app.
- Follow the in-app instructions to download a local model: PPLX Pro or Qwen 3.6.
- Choose fully local or Hybrid for each task.
The Verge summed up the catch in its headline: you’ll need a lot of RAM. Many base-model Macs ship with far less than 36GB, so this is aimed at higher-end machines. If you are weighing a Mac purchase, our explainer on unified memory vs SSD storage explains why memory, not storage, is the number that matters here.
Why it matters
Running an AI agent locally changes three things:
- Privacy: confidential files do not need to leave the machine for the AI to work on them, which matters for finance, legal and client work.
- Cost control: Perplexity says local mode puts the choice of where tasks run, and how credits are spent, in the user’s hands.
- Offline work: tasks that do not need live data can keep going without a connection.
The trade-off is capability. A model small enough to run on a laptop is generally less capable than the largest cloud models, which is presumably why Perplexity keeps Hybrid Compute and the option to escalate. How well local Computer handles complex, multi-step work will become clearer as people test it on real tasks.
The bottom line
Perplexity Computer local on Mac is one of the clearest signs yet that AI agents are moving onto personal hardware. If you have an Apple silicon Mac with plenty of memory and a paid Perplexity plan, you can now run agent tasks on private files without those files reaching the cloud. For everyone else, the 36GB recommendation is the hurdle. For more on Perplexity’s recent moves, see our coverage of its new embedding models.
The Teck Hustlers Editorial Team covers technology, business, finance and gaming news for readers who want the story behind the headline. Every article is researched, written and edited by our in-house team, checked against company announcements, filings and reputable reporting, and updated when new facts emerge. We focus on what each development means for founders, professionals, investors and everyday users. Have a tip or a correction? Reach us through our contact page.
