Description
Models like DeepSeek or Qwen need more VRAM than one machine has, and cloud clusters are expensive. Parallax is a distributed model serving framework that turns your devices into your own AI cluster.
It supports decentralized inference, mainstream models like Llama, Qwen, GLM, Kimi and MiniMax, Blackwell GPUs and OpenClaw.
Distributed:Many machines, one model.
Decentralized:Devices anywhere.
Models:Qwen, DeepSeek, GLM and more.
OpenClaw:Integrated.
It supports decentralized inference, mainstream models like Llama, Qwen, GLM, Kimi and MiniMax, Blackwell GPUs and OpenClaw.
Features
Distributed:Many machines, one model.
Decentralized:Devices anywhere.
Models:Qwen, DeepSeek, GLM and more.
OpenClaw:Integrated.
