GoKawiilPerplexity debuts on-device AI agent 'Portable Computer' with NVIDIA
Perplexity has released Portable Computer, an agentic AI tool that runs entirely on local hardware instead of the cloud, built in partnership with NVIDIA. It relies on compact models like Qwen 3.8 and NVIDIA Nemotron 3.5 Lightning to handle complex tasks on-device, and only reaches out to cloud-based models when needed, with user permission. The feature is currently available to Perplexity Pro and Max subscribers running NVIDIA's DGX Spark hardware, with RTX GPU support planned.
GoKawiil's interpretation of the reporting above, not reported fact.
By keeping processing local, Perplexity avoids sending sensitive data to external servers and eliminates token costs for on-device tasks, addressing both privacy and pricing concerns that have dogged cloud-based AI agents. This positions Perplexity to compete for users who need high-volume or privacy-sensitive automation without racking up usage fees, while NVIDIA gains another showcase for its DGX Spark and edge AI hardware.
- Portable Computer runs locally using small models like Qwen 3.8 and NVIDIA Nemotron 3.5 Lightning.
- Tasks completed on-device incur no token costs, though cloud features require user approval.
- Currently limited to Perplexity Pro/Max subscribers on NVIDIA DGX Spark, with RTX GPU support coming soon.
NVIDIA GeForce RTX 4060 Graphics Card — Running local AI agents like Perplexity's Portable Computer depends on having a capable GPU in your machine. An RTX 4060 gives you NVIDIA hardware acceleration for on-device models, so your data stays local instead of going to the cloud. It's a practical upgrade for anyone who wants to experiment with local LLMs and agentic tools.
See NVIDIA GeForce RTX 4060 Graphics Card on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: androidauthority.com, 2026-08-26
Published there as: “Perplexity launches a local AI agent with zero token costs”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.