Perplexity Computer now runs entirely on Apple Silicon Macs, but its recommended RAM requirement is steep and limits real-world deployments. The fully local mode launched October 20 for Perplexity Pro, Max, and Enterprise subscribers. It builds on April's Personal Computer launch and September's Hybrid Compute update. Personal Computer brought AI automation to native Mac apps, while Hybrid Compute paired local processing with cloud-based reasoning.
Users can now choose to keep an entire task on their Mac or authorize cloud processing when additional capabilities are needed, assuming the minimum requirements are met. The fully local version supports Qwen 3.6 and PPLX Pro, a Perplexity-trained version of Qwen. Both models run on the Mac, handling reasoning and coordinating tasks without relying on remote AI servers.
With permission, Computer can work across native apps including Notes, Messages, Mail, and Numbers. It can review messages and emails, identify outstanding deadlines, and create a task list in Notes. It can also analyze private financial information against downloaded market data, update an Excel workbook, and prepare a Keynote presentation. Perplexity says code execution takes place in isolated sandbox environments with controlled access to files and supported apps.
Fully local processing doesn't make every workflow independent of the internet. Sending an email or accessing an online service still requires a network connection. The bigger hurdle is memory: Perplexity recommends at least 36 GB of unified memory, which exceeds the maximum available in some recent Apple silicon Macs. The standard M6 Mac mini tops out at 32 GB, while the M5 Pro Mac mini can meet the recommendation with 48 GB or 64 GB configurations.
Source: AppleInsider · Summarized by HeadlinesBriefing