Capabilities
One AI layer. Everywhere you work.
TARX brings private AI across your computer, files, memory, tools, agents, devices, and Supercomputer power - starting free and local.
Local AI
Start on your computer. Run private AI locally before sending work anywhere else.
Memory
Remember what matters. Store user-approved context: projects, preferences, decisions, workflows, and operating assumptions.
Vault
Keep secrets out of memory. Credentials, API keys, tokens, SSH keys, private keys, passwords, passcodes, and sensitive secrets belong in Vault.
Agents
Move from answers to action. TARX can support agents that use tools, context, and approvals to complete workflows.
Skills
Give AI specific abilities. Skills package repeatable work so TARX can operate across domains, tools, and user needs.
Tools
Connect to real work. Use APIs, files, local services, integrations, and approved actions.
Desktop
AI where the work already lives. TARX should operate on the computer, not only inside a browser tab.
Web
A surface for access, account, docs, and Supercomputer.
iOS
Private AI should move with you.
On-device AI
Use on-device intelligence to improve privacy, responsiveness, and personalization where appropriate.
Supercomputer
Power when local is not enough. Route approved heavy work to Supercomputer. Charge for power, not base access.
Mesh compute
Give compute. Get compute. Contribute useful compute to earn access back. Buy power when you use more than you give.
Enterprise Runtime
Governed AI for organizations. Private deployment, policy, evidence, routing, Vault boundaries, procurement, support, and deployment control.
Operational integrations
Built for serious environments. TARX can support operational AI use cases, including Palantir-adjacent environments where context, tools, evidence, and governance matter.
