NVIDIA has released the open-source beta version of the Personal AI Router (PAIR) software, which intelligently schedules idle devices for local AI requests within a local area network. It is compatible with Mac devices (M4 and above), NVIDIA RTX PCs, and DGX Spark devices, providing a unified entry point for local AI tools such as Ollama and LM Studio. PAIR can automatically discover devices, allocate requests, while the model execution remains the responsibility of the tools on the target devices, eliminating the need for separate configuration of each machine for every AI application. When a device is busy or unavailable, the software will redirect requests to share concurrent loads, though the speed of individual AI responses will not automatically increase with more devices. Demonstrations show that multi-device collaboration can reduce multitasking processing time.
