How to Use Perplexity Portable Computer

The Short Answer

Use Perplexity Portable Computer on Windows, or fully local Computer in the Mac app, this way:

1. Check the platform. AMD Windows needs Pro or Max on an individual or enterprise plan, Windows 10 or 11, a Ryzen AI Max Series system with at least 24 GB of GPU-accessible memory, and roughly 20 GB of free storage. Mac needs a supported Apple silicon machine, the Perplexity Mac app on macOS 15 or later, and Pro, Max, or Enterprise; 36 GB or more is recommended for fully local work.
2. Install the current Perplexity desktop app for your platform and sign in.
3. Download a local model. On Windows, use Settings > Local Inference for PPLX 27B or Qwen 27B. On Mac, follow the app's setup instructions for PPLX Pro or Qwen 3.6.
4. Grant file permissions and select one test folder. Windows uses Settings > Permissions; Mac uses Perplexity > Settings… > Permissions in the macOS menu bar. In Computer, select the task's folder through Folders > Choose folder.
5. Choose local execution. On Windows, select the downloaded model under Local in the composer. On Mac, choose fully local; Hybrid starts with cloud work.
6. Run a small file task with known answers before you schedule anything important.

Since October 20, 2026, Computer can run fully local on supported Apple silicon Macs, alongside the AMD Windows route introduced September 24. Requirements and setup were verified against Perplexity's own pages on October 22, 2026.

Portable Computer Is the Local Brain, Not Just File Access

Personal Computer's file tools and local inference solve different halves of the job. Think of file access as the key to a filing cabinet: it lets Computer find, read, create, and edit local files. Local inference moves the analyst into the room: the agent and model also run on your device.

You can use local file tools without turning on Portable Computer. You can also use both together, which is the setup this guide uses. On Windows, Perplexity says the local model, planner, tool router, scheduler, durable task queue, file search index, and sandbox run on the PC. The Portable Computer help page explains the distinction.

On Mac, fully local keeps planning, reasoning, and file processing on the Mac. Hybrid Compute starts in the cloud and delegates private-file work to a local model. Downloading a local model alone does not make a Hybrid task fully local. Perplexity's Mac app page shows both modes.

That boundary matters because a local task can still ask for cloud help with current information, browser use, connected apps, or harder reasoning through more than 15 frontier models. Users can authorize that escalation, as the Mac announcement explains. Work handled by the local model uses no Computer credits on either platform; cloud steps use credits like other Computer work. Perplexity's help center states the credit rule. Fully local describes where the AI runs: sending a message or meeting invitation still uses the relevant app's service, as the Mac product page clarifies.

Check the Machine Before You Download Anything

Check the requirements for the mode you intend to use. A Windows PC with 24 GB of ordinary system RAM does not automatically qualify, and the Mac's 24 GB Hybrid requirement is different from its fully local recommendation.

The AMD requirements are unchanged. The Mac launch announcement supplies the fully local recommendation and model choices; the Mac app download page supplies the operating-system requirement. Perplexity's Hybrid Compute announcement documents the lower memory requirement for that separate mode. Treat 36 GB as a recommendation, not a published hard minimum.

There is a documentation mismatch worth seeing before it wastes an afternoon. The older Portable Computer help article still lists DGX Spark or supported NVIDIA RTX hardware. The AMD announcement and Personal Computer help page explicitly include Ryzen AI Max. That help page still describes Mac's older Hybrid mode, so use the newer Mac announcement and current Mac product page for fully local model choices.

You still need a paid Perplexity plan. Its consumer pricing page displays Pro at $17 per month and Max at $167 per month when billed annually. Hardware, electricity, review time, and any cloud work remain costs.

The useful budget equation is:

Monthly local workflow cost = paid plan + hardware amortization + review time + cloud credits

If you already own a qualifying machine and already have an eligible plan, the extra Computer-credit cost of a fully local run is zero. If you need new hardware, do not force a payback claim from an unspecified purchase price. Measure the job first.

Set It Up Without Widening the Permission Boundary

The safest first setup is one model, one folder, and one task with answers you already know.

1. Install the app for your platform and sign in

On Windows, install the Perplexity app. Portable Computer is not a second Windows installer: local inference is a setting inside the app.

On Mac, download the current app from Perplexity's Mac page, open the disk image, move Perplexity to Applications, then open it and sign in. Check for updates through Perplexity > Check for Updates… in the macOS menu bar. These are the documented desktop setup steps.

2. Download one supported local model

Windows: Open Menu > Settings > Local Inference. Under On-device models, download PPLX 27B or Qwen 27B. The inference engine and sensitive-content classifier install with the model. Only one local model runs at a time on Windows, so choose one for the first test.

Mac: Follow the in-app instructions to download PPLX Pro or Qwen 3.6. Perplexity's current Mac setup page documents that sequence without naming a Local Inference submenu. Use the app's instructions for the download.

3. Create one fixture folder

Create a folder called Portable-Reconcile-Test. Put only synthetic test files in it. Do not start with a live inbox, payroll export, contract repository, or shared drive.

4. Permit only that folder

Windows: Open Settings > Permissions. Enable Local file system access, then turn on Desktop tools.

Mac: Open Perplexity > Settings… > Permissions from the top-left macOS menu bar, or choose Settings… from the Perplexity menu-bar icon. Approve the permissions needed for the task. Mac file access uses Full Disk Access, folder prompts, and the Filesystem connector. There is no Windows-style Local file system access row or Desktop tools switch on Mac.

Both platforms: In Computer, open Folders > Choose folder and select only Portable-Reconcile-Test for this task. Perplexity documents folder selection per task, separately from the app's permission settings. A smaller permission boundary makes a bad instruction easier to detect and reverse. Desktop permissions and folder setup.

5. Select local execution before the task

Windows: Return to the composer, open the model menu, and select the downloaded model under Local. A cloud model selected in the composer is a different execution path.

Mac: Choose fully local, as shown in Perplexity's Mac setup instructions. Choose Hybrid only when you intend to combine cloud work with local file processing. For this fixture, keep the task fully local.

6. Decline cloud escalation during the fixture run

The test below does not need current web data, a connector, or frontier reasoning. If an approval prompt appears, decline it and record why the task asked. A clean local test should not require a cloud step.

Run a Reconciliation With Three Known Exceptions

This fixture is designed to fail in exactly three known ways. It gives you something better than a vague demo: a result you can mark right or wrong.

Create contract-rates.csv with these four rows:

vendor,lane,contract_amount
Northstar Freight,ATL-MIA,1320
Blue Harbor Logistics,DAL-HOU,930
Juniper Transit,BOS-NYC,660
Cedar Route,SEA-PDX,550

Create invoices.csv with these five rows:

invoice_id,vendor,lane,invoice_amount
INV-1001,Northstar Freight,ATL-MIA,1320
INV-1002,Blue Harbor Logistics,DAL-HOU,980
INV-1003,Juniper Transit,BOS-NYC,660
INV-1003,Juniper Transit,BOS-NYC,660
INV-1004,Maple Freight,DEN-PHX,745

Then give the local model this instruction:

The expected exception file has three findings:

invoice_id,issue,expected_amount,invoice_amount,difference
INV-1002,RATE_VARIANCE,930,980,50
INV-1003,DUPLICATE,660,660,0
INV-1004,NO_RATE,,745,

Check the row count, each invoice ID, and each amount. Do not score the run by how polished the explanation sounds. The output is correct only if it catches those three exceptions without adding a fourth.

This procedure is reproducible, but it was not executed in this article. It is an acceptance test you can run on either platform, not a measured result from a supported PC or Mac.

Schedule the Same Task, Then Keep the Machine Available

Schedule the reconciliation only after the one-off run passes. Reuse the same folder, same instruction, and same expected schema. On Windows, Perplexity documents a local scheduler and durable task queue; scheduled work runs while the app stays open and the PC stays awake. AMD recurring-work requirements.

On Mac, keep the selected machine available, awake, online, and signed in, with the app running, as Perplexity's Mac availability guide requires for desktop work. The Mac launch pages do not document a fully local scheduling sequence. The separate Scheduled Tasks guide describes cloud execution and credit use, so its Automations controls are not evidence of a fully local Mac schedule. Keep the Mac pilot manual until you can verify where its recurring runs execute.

That availability condition is operational, not cosmetic. A sleeping laptop is not a dependable morning control. If the exception file matters to a close process, use a machine with an explicit power policy, name an owner for failed runs, and keep the manual check until the schedule has proved reliable.

Record these fields for the first real test:

• Hardware: Exact Ryzen AI Max system and GPU-accessible memory, or Apple silicon Mac and unified memory.
• Software: Operating system, Perplexity app, local model, and execution mode.
• Setup: Reported download size and measured setup time.
• Run: Start time, finish time, and whether the app stayed open.
• Output: Three expected findings, extras, and misses.
• Credits: Visible balance before and after.
• Boundary: Every cloud approval prompt and your response.

An unchanged credit balance would support the narrower claim that the run did not consume Computer credits. It would not prove zero network activity. If you need that assurance, use network monitoring and an approved security test rather than treating the product label as evidence.

Six Workflows That Benefit Most

The best use cases are repetitive, file-heavy jobs where a human can review a compact exception output. These are workflow ideas, not measured results; Perplexity documents the AMD examples and Mac file-and-app capabilities.

The common shape is exception-first automation. Portable Computer is less attractive when every task needs live web data, the machine cannot stay awake, or nobody owns the review step.

Three Things Worth Building

1. A local invoice-exception runner

This is the strongest opportunity. Build a small deployment package for controllers and bookkeeping firms that installs the folder structure, validates CSV columns, runs the reconciliation prompt, writes a stable exception schema, and keeps a run log.

The smallest sellable version needs three things: a fixture-based acceptance test, reusable mappings for two CSV exports, and a human review screen for exceptions. The catch is accounting variability. Column names, currencies, partial payments, and approval rules differ by customer, so a generic prompt is not the product. The mappings and evidence trail are.

2. A private contract-clause checker

Package a permitted-folder workflow for legal operations teams: one clause checklist in, a contract set in, and a cited exceptions table out. The buyer pays for faster issue spotting, not autonomous legal judgment.

An MVP can support one contract type, one approved checklist, clause-level source references, and mandatory lawyer review. The catch is decisive: a missed clause can cost far more than the software. Without strong evaluation sets and a visible review trail, this should not ship as a decision maker.

3. A local-compute readiness and runbook kit

Build a paid setup service or lightweight verifier for IT teams adopting local agents on Ryzen AI Max PCs or Apple silicon Macs. It would check the platform-specific processor and memory requirements, download space, app version, local model state, execution mode, folder permissions, power policy, and a fixture result.

The MVP is a readiness report plus the exact three-exception test in this guide. The catch is platform risk. Perplexity can improve its own setup diagnostics. This works better as a wedge into local-agent implementation than as a standalone SaaS bet.

Where Portable Computer Stops Being the Right Choice

Do not use it just because “local” sounds safer. The hardware must qualify, the plan is paid, and local work needs an available machine with the app running. The Windows Portable Computer guide limits execution to one local model at a time, excludes Brain and saved memory from local mode, and describes the local model as smaller than a frontier model. Those are documented Windows limitations, not a separate list of confirmed Mac restrictions.

The security boundary also needs precise language. Portable Computer controls folder access and sandboxes code execution, but a local MCP server runs outside that sandbox with your user permissions. Connected apps, browser use, web research, and approved cloud reasoning can cross the device boundary. On Mac, sending a message still uses the app's service even when the AI runs locally. These controls do not prove that the machine makes no network calls. Mac processing and service boundaries.

Choose another route when you need an air-gapped guarantee, an always-on shared server, a multi-user policy layer, unsupported hardware, or a workflow that constantly depends on the web. For the wider product decision, see the Portable Computer alternatives comparison. Use the Perplexity sources above for current hardware eligibility. The broader Perplexity review is the better branch when your question is whether Perplexity fits your research work at all.

The Monday Move

If you already have a qualifying Ryzen AI Max Windows PC or Apple silicon Mac and an eligible plan, reserve a focused session next week for one bounded pilot. Permit only Portable-Reconcile-Test, download one model, run the three-exception fixture, and record the result. Schedule it only after the output is exact and you have verified the recurring execution mode. If the machine fails the eligibility check, stop there. Do not buy hardware until a real recurring workload and its current labor cost justify the budget.

If you want a bounded local workflow like this built and tested for your business, I can help with the automation design.