THE CRUNCH
Perplexity has released Portable Computer for Windows, a local AI agent that runs on RTX GPUs with at least 24GB of VRAM. The app, available via the existing Perplexity client, brings multistep task planning and local model support to compatible PCs. Users must have a GeForce or RTX PRO GPU with 24GB or more of video memory and a Pro or Max subscription. The agent can perform tasks locally without consuming credits,,
The Windows release adds scheduled recurring tasks and local MCP servers for desktop apps, such as Microsoft Word, Google Drive, and Slack. Nvidia has integrated connectors for these services into the app, and users can download a local model like Qwen 3.8 27B with a single click. The 24GB VRAM requirement spans Nvidia’s consumer lineup, including the RTX 3090, 4090, and the upcoming 5090, as well as RTX PRO Blackwell cards. This lowers the barrier to entry for fully local AI compared to the standalone DGX Spark hardware, which costs $4,699.
Perplexity CEO Aravind Srinivas described the release as bringing “unmetered local intelligence on every Windows PC running on Nvidia hardware and Perplexity harness.” The company previously offered this functionality only on Linux. Nvidia also announced that RTX Spark Windows PCs from Lenovo and Acer are expected in October, alongside simplified local setups for other agents like Hermes and OpenClaw.