Office

Separate model services keep coding work and everyday business requests from competing for the same machine. Office is designed for a mixed team of up to eight people. That figure is the intended mix of users, not an unlimited number of simultaneous generations, and it is subject to workload validation.

The team can code with a local agent, search shared documents, and draft from private files. Team software adds user access, request routing, workload queues, and monitoring. We size the system around how many people generate responses at once, their context needs, and how heavily they use coding agents.

Supported distributed software can split a model across the two Macs. Each Mac still has its own memory, and distributed execution has communication overhead. For many office workloads, running separate model services offers a better experience.

Our custom enclosure organizes the computers, cabling, and airflow, and keeps each node accessible for service. The nodes connect over Thunderbolt 5 and your local network, with 10 GbE networking and 4 TB of external storage for documents.

Memory is not pooled. Each Mac Studio has 128 GB of unified memory of its own. Multi-Mac systems can expand later with additional qualified nodes. We document the supported upgrades with your configuration.

Apple Silicon inference uses MLX and compatible local runtimes, configured for these machines.