Local-first
Your agents. Your machine. Your data.
Matrix is built from the ground up for local deployment. Run the full stack — Executor, Cortex, UWAC, and your agents — entirely on your device, with zero cloud dependency.
True privacy
No requests leave your machine. Your context, memory, and credentials stay local and are never transmitted to Matrix servers.
Offline capable
The Executor and Cortex run fully offline. When connectivity returns, cloud sync is opt-in and scoped.
Model-agnostic
Run with any local model via Ollama, LM Studio, or a custom inference endpoint. No API keys required.
Local by design, not by compromise
Most “local” AI products are cloud products with an offline mode bolted on. Matrix is different. Every component — the Executor daemon, the Cortex Pebble store, UWAC credential vaulting — was designed to run on a single machine first. The cloud version of Matrix is the distributed extension of the local version, not the other way around.