Nvidia Announces PAIR, a Personal AI Tool That Combines Idle Computers
Nvidia has introduced PAIR, a free, open-source software that combines idle computers at home for local artificial intelligence processing.
Nvidia has announced Personal AI Router (PAIR), a free and open-source software that connects idle computers at home over a network to run local AI and agent workflows in parallel.
The Core Purpose of the Software
Developed by Nvidia, the Personal AI Router turns the power of powerful, idle computers at home into a single AI data center.
This system is designed to perform local AI inference tasks using tools like Ollama and LM Studio.
Supported Hardware
PAIR is not a hardware-based router; it is open-source software that discovers and connects compatible computers on the network.
Compatible devices include Nvidia GeForce graphics cards of the RTX 20 series and above, RTX Pro, DGX Spark systems, and devices equipped with Apple M4 or newer chips.
Parallel Operation and Flexibility
The system activates when home devices are idle, operating in the background without interfering with other daily operations.
The software dynamically adapts as devices are added to or removed from the network; for example, the system can adapt the flow even when gaming starts on a desktop computer.
Security Measures
Nvidia states that all devices are secured after being paired with a six-digit code.
The communication channel is encrypted using Mutual Transport Layer Security (mTLS) technology, which ensures mutual trust between computers.
Beta Version and Application Support
A beta version of the PAIR software, supporting Windows, Linux, and macOS operating systems, is now available.
Additionally, three major free AI agent applications—Perplexity Portable Computer, Hermes Agent, and OpenClaw—will offer easy installation support with Nvidia GPUs on Windows.