EXO1.0.71

EXO connects Apple Silicon Macs into a local cluster so you can run AI models across multiple devices that you own. It distributes a model across the connected machines, allowing larger models to use their combined resources while optimizing runtime parameters automatically. Local inference keeps requests on your own hardware, whether that means one computer or a cluster on your local network.

Connect devices over Thunderbolt or USB-C, then view the cluster and manage its models from a single dashboard. EXO runs open models through the harness you already prefer and exposes familiar integration formats, including OpenAI Chat Completions, OpenAI Responses, Claude Messages, and Ollama. Developers can use it to test or deploy local AI while retaining control of the machines and infrastructure handling their models.

The software suits individual developers, teams exploring private local AI, and anyone assembling a multi-device setup for models that exceed one machine’s practical capacity. Its open-source code can be inspected, improved, or extended, and the system is designed to make connected devices operate as one local AI environment. Version 1.0.71 runs on macOS 15 Sequoia or later and requires Apple Silicon (arm64) hardware. EXO is freeware in the Developer Tools category.