TX-LOCAL
Runs entirely on your hardware
One command installs TARX. The Bridge starts, loads a quantized model, and your first inference runs in seconds. No account required. No API key. No GPU rental. Your Mac is the supercomputer.
Now available for macOS + iOS
Local. Private. Proactive.
AI that runs on your hardware. Never sends your data anywhere. TARX is the supercomputer that lives on your Mac — private by architecture, always on, always learning how you think.
curl -fsSL tarx.com/install | shmacOS · iOS · No account required
Why TARX
PRIVATE
Your data never leaves your device
TARX is private by architecture, not by policy. Inference happens locally. Memory stays on-device. There is no cloud to breach because your data never leaves your Mac. Zero telemetry. Zero training on your content.
SUPERCOMPUTER
Scale beyond one machine when you need it
Connect multiple Macs into a single mesh. TARX distributes inference across your own hardware — your studio, your office, your fleet. Enterprise-grade AI without sending a single byte to someone else's server.
How TARX works
TARX is a local-first AI system. It runs a fine-tuned language model directly on Apple Silicon, learns how you work, and surfaces what matters before you ask.
1
Install in seconds
Download the app or run the install script. TARX loads a quantized model optimized for your Mac. No configuration needed.
2
It learns you
TARX adapts to your role and reasoning style. A Soul Layer fine-tunes locally over time so responses get sharper the more you use it.
3
Always on, always private
TARX runs continuously in the background. It surfaces relevant news, notices patterns, and builds automations — all without ever phoning home.
Free to think.
TARX is free for personal use. Download for Mac, install in one command, and start using AI that respects your privacy and runs on your terms.
Free personal tier · $6/seat/mo for teams · No credit card required
TARX IS REASONING▾