Home / Technology / Perplexity AI: Run Fully On Your Own Device
Perplexity AI: Run Fully On Your Own Device
20 Oct
Summary
- Users can now run the entire Perplexity AI experience locally on their Macs.
- Requires a recent Apple Silicon Mac with at least 36GB of RAM.
- Offers enhanced data privacy and control over AI workload spending.
Perplexity AI has launched a significant update enabling users to run its entire platform locally on their personal devices. This feature is currently available for users with recent Apple Silicon Macs equipped with at least 36GB of RAM.
The capability extends to running the core harness and orchestrator on the user's machine. This move is pitched as beneficial for privacy-conscious individuals and those seeking to manage their AI expenditure more effectively. For instance, local execution allows the AI to process sensitive data from Mail or Notes without cloud transmission.
Users can opt for two primary models for local operation: Qwen 3.6 or PPLX Pro, a fine-tuned Perplexity version of Qwen 3.6. The system also permits hybrid operation, allowing tasks to utilize over 15 cloud-based models if needed. All code and tool executions are performed within a secure, isolated sandbox environment.
This local compute functionality is available immediately for Perplexity Pro, Max, and Enterprise subscribers. Installation involves downloading the latest Perplexity Mac app and following in-app directions to set up a local model.
