Perplexity Computer Now Runs Entirely on Your Mac with Full Local AI Processing

Reviewed byNidhi Govil

7 Sources

Share

Perplexity has launched a fully local version of its Computer app for Mac, allowing AI workflows to run entirely on-device without cloud dependency. Available to Pro, Max and Enterprise subscribers, the update requires Apple Silicon Macs with at least 36GB of RAM and supports Qwen 3.6 and PPLX Pro models for private, offline AI computing.

News article

Perplexity Computer Launches Fully Local AI Mode for Mac

Perplexity has rolled out a significant update to its Computer Mac app, enabling users to run AI workflows entirely on their own device without relying on cloud infrastructure

1

. This local-first mode represents a shift toward on-device AI processing, allowing the entire Computer experience—including the harness and orchestrator—to operate locally on Apple Silicon Macs

2

. The feature is available now to Perplexity Pro, Max and Enterprise subscribers who own recent Apple Silicon Macs with at least 36GB of RAM

3

.

The update builds on Hybrid Compute, introduced in September, which split AI workloads between cloud-based frontier models and local models running on users' machines

4

. The new fully local option takes this further by handling AI model execution, task planning, reasoning, and file processing entirely on-device

5

. Perplexity highlights the new Mac mini machines as compatible devices and recommends at least 36GB of memory for optimal local use

3

.

Data Privacy and Offline AI Computing Drive Adoption

Perplexity is positioning this feature as a solution for users prioritizing data privacy and cost control. With access to native Mac apps like Notes, Mail, Messages, and Numbers, Perplexity Computer can generate priority to-do lists, review deadlines, and organize meetings without sending personal data to the cloud

1

. For instance, someone reviewing an investment portfolio can use locally stored holdings and market data, then bring in Hybrid Compute only when fresh market research is needed

5

.

The ability to run tasks without an internet connection marks a notable advancement in offline AI computing

2

. Files only reach the cloud when users explicitly approve, giving them granular control over sensitive information

2

. Perplexity notes that "as more demanding workflows become practical to run locally, the machines people already own will take on a larger share of their AI work," offering users more control over sensitive data and credit spend

4

.

Qwen 3.6 and PPLX Pro Power Local AI Workflows

Users can choose between two models to run Perplexity Computer locally: Qwen 3.6 or PPLX Pro

1

. PPLX Pro is Perplexity's post-trained version of Qwen 3.6, specifically fine-tuned for Computer use through additional training

3

. Both models can be downloaded and installed directly through the Perplexity Mac app following in-app instructions

1

.

When a task demands more advanced reasoning or current information, users can authorize the software to call on more than 15 frontier models for cloud-based assistance

4

. This flexible approach lets users decide when to spend credits and when to keep processing local

2

. Any code or tool execution required by an agentic task runs in an isolated sandbox environment, with Perplexity confirming that "these same security protocols persist in the cloud version of Perplexity Computer"

1

.

What This Means for Mac Users and AI Adoption

This update matters because it addresses two critical concerns holding back AI adoption: data privacy and infrastructure costs. By enabling run entirely on your Mac workflows, Perplexity removes the need to send sensitive business documents, personal emails, or confidential notes to external servers. Users can draft client proposals using confidential pricing information or sort through messages to identify deadlines without exposing private data

5

.

The hardware requirements suggest this is aimed at power users and professionals willing to invest in capable machines. While Hybrid Compute works on Macs with as little as 24GB of memory, the fully local option requires at least 36GB of RAM, limiting accessibility to higher-end configurations

3

. Watch for how this influences Mac hardware purchasing decisions among AI-focused professionals and whether Apple responds with memory configuration adjustments in future releases. The shift toward local AI also raises questions about how quickly consumer hardware will need to evolve to support increasingly demanding on-device AI processing workloads.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved