Skip to content
All tool news

Tool desk · Updated daily

Tool news·Perplexity·

Perplexity Portable Computer Brings Local AI Agents to DGX Spark.

Perplexity Portable Computer runs its agent stack locally on NVIDIA DGX Spark, with user-approved cloud access for harder tasks.

CW

Create With tool desk

1 source checked · 3 min read · 3 sections

ShareLinkedIn

Perplexity is bringing its Computer agent off the cloud and onto local hardware. The company announced on LinkedIn that Portable Computer runs the full agent runtime on NVIDIA DGX Spark, allowing eligible users to work with sensitive documents without sending those files to a cloud service.

The release covers more than local model inference. Perplexity says the orchestrator model, subagent model and agent harness all execute on the DGX Spark. That makes Portable Computer a locally operated version of Perplexity Computer rather than a thin desktop interface that still depends on remote infrastructure.

Portable Computer is available as of August 25, 2026, to Perplexity Pro and Max subscribers who have an NVIDIA DGX Spark.

A Local Runtime With an Optional Cloud Route

Portable Computer uses PPLX 27B, a post-trained model installed locally. Perplexity is also making Qwen 3.8 27B available as an alternative, while support for NVIDIA Nemotron 3.5 Lightning is listed as coming soon. The announcement does not provide a date for that additional model.

The local-first design does not completely remove access to cloud models. If a task requires what Perplexity calls “frontier reasoning,” the orchestrator can request permission to route the task to a frontier model in the cloud. The request is user-gated, so cloud processing should not begin without approval.

Perplexity also describes the handoff as PII-flagged and limited to text guidance. Sensitive documents remain on the device, according to the company. The announcement does not detail exactly how personally identifiable information is detected, what text is included in an approved request, or whether administrators can impose broader organization-level restrictions.

The resulting workflow has three notable characteristics:

  • Local execution: The orchestrator, subagents and agent harness run on the DGX Spark.
  • Explicit cloud approval: Harder tasks can be escalated only after the user approves the request.
  • Document separation: Perplexity says sensitive source documents stay on the local device, even when cloud guidance is requested.

Why the Architecture Matters

For small businesses, agencies and professional teams, agentic AI creates a difficult trade-off. An agent becomes more useful when it can inspect contracts, internal research, customer records or financial documents, but those are also the materials teams may be least willing to upload to a third-party cloud.

Portable Computer is Perplexity’s attempt to separate document access from frontier-model reasoning. Routine work and document handling can remain local, while a cloud model can provide limited assistance when the installed model is not sufficient. The approval step also gives the operator a decision point that is often missing from fully cloud-based agent workflows.

There are practical constraints. DGX Spark is specialized AI hardware, so this is not a local mode intended for an ordinary business laptop. Access also requires a Pro or Max subscription. Perplexity has not disclosed separate Portable Computer pricing, minimum storage requirements, model download sizes or performance benchmarks in the supplied announcement.

The launch also broadens the company’s approach to Computer. Other recent updates have concentrated on bringing more information and models into cloud workflows, including licensed business-data connectors and U.S.-hosted DeepSeek V4 Pro. Portable Computer moves in the other direction, putting the execution environment beside the user’s data and making cloud access an optional, visible step.

For organizations evaluating it, the key question will be whether local PPLX 27B or Qwen 3.8 27B can handle enough day-to-day tasks without frequent escalation. Perplexity has established the architecture, but real-world benchmarks and more detailed controls will determine how useful that architecture is for sensitive production work.

Frequently asked questions

What is Perplexity Portable Computer?

Perplexity Portable Computer is a locally running version of Perplexity Computer for NVIDIA DGX Spark. Its orchestrator model, subagent model and agent harness run on the user’s hardware rather than depending on Perplexity’s cloud runtime.

Can Portable Computer use cloud AI models?

Yes, but only as an approved fallback for tasks requiring frontier reasoning. Perplexity says the system asks the user before routing text guidance to a cloud model, flags personally identifiable information and keeps sensitive documents on the device.

Who can use Portable Computer, and which models does it support?

It is available to Perplexity Pro and Max subscribers with an NVIDIA DGX Spark. It runs a locally installed PPLX 27B model, also supports Qwen 3.8 27B, and is expected to add NVIDIA Nemotron 3.5 Lightning support later.

Sources

1 checked

How we cover tool news: Create With's tool desk drafts these reports with AI from the sources listed above and checks them against those sources before publishing.

Worth passing on?

ShareLinkedIn

Go deeper on Perplexity

Related reading, watching and going.

Everything on Perplexity →

The briefing

16 Jun 2026

How Claude Code Lets One Person Replace SEO Teams Fast

Discover how Claude Code lets one non-technical user automate SEO — from keyword discovery to publishing — replacing SEO teams easily without writing code.

Read →

The briefing

31 Mar 2026

Vibe Marketing Review: Can AI Agents Really Automate Marketing?

An honest review of vibe marketing with AI agents. Learn if automated marketing actually works for solo founders and small teams. Plus: Claude Code leak, Factory Missions.

Read →

The briefing

3 Mar 2026

Perplexity Computer Platform & Notion AI Agents Review 2025

Perplexity's 19-model AI platform, Claude Code mobile control, iPhone offline AI with Qwen 3.5, and Notion's custom agents are transforming AI automation.

Read →
31 min

Conference

How to Get Your Product Recommended by Google and ChatGPT

Tanya Van Gastel presents the APEX framework for making products discoverable in AI-powered search engines like ChatGPT and Google's AI Overviews. She explains how AI is fundamentally changing product discovery and buyer behavior, then breaks down five critical components: Authority (being cited as a solution), Presence (visibility across search platforms), Expertise (creating high-quality money pages), Endorsements (off-site validation), and Extractability (making content easy for AI to process). The talk emphasizes quality over quantity, with data showing 50% of SaaS SEO revenue comes from fewer than 10 high-performing pages.

Watch now ↗

Latest tool news

What else changed this week.

All tool news
VercelDigest

What Vercel Shipped in Late October

Vercel shipped quicker static deployments, four AI Gateway models, more free domain choices and new CLI controls in late October. The changes cover developers testing sites…

Vercel

Vercel Adds Submit_feedback to Its MCP

submitfeedback is a new Vercel MCP tool that lets agents and harnesses send bug reports, friction reports and feature requests directly to the Vercel team. Feedback can include…

The Create With Briefing

Don't watch forty changelogs. Read one email.

Every Tuesday: the tool changes worth knowing, real business use cases, and what's on near you. Free, unsubscribe any time.