01
Local work
Useful AI work can start against local files, local context, local tools, and user-approved actions.
FREE TO THINK
● PRIVATE PREVIEW UTC
● LOCAL-FIRST AI APPS
Build AI apps
that start local.
TARX lets your app start on the user's machine, scale to hosted compute, and move into enterprise-owned infrastructure without changing the runtime contract.
Local runtimeHosted headroomSame contractEnterprise path
Why local-first matters
The app starts
where the work lives.
AI apps get harder to govern when every action starts in a remote cloud. TARX gives builders a runtime posture that begins on the user's machine, keeps local work visible, and adds hosted Supercomputer headroom only when the job needs more power.
Runtime shape
What runs
locally first.
Local-first does not mean pretending hosted compute is useless. It means the first execution surface is the computer the user already owns.
02
Visible state
The runtime should show what is running, what evidence exists, and what needs approval.
03
User control
Local-first apps can keep approval, memory, and skill boundaries closer to the user.
When local is not enough
Add hosted
Supercomputer headroom.
TARX keeps the contract steady as work moves from local runtime to hosted compute. Paid plans should read as more hosted Supercomputer headroom, not a different product.
01
Burst work
Heavier jobs can use hosted headroom when local compute is not the right fit.
02
Same runtime contract
The app should not need to rebuild its user model every time compute changes.
03
Enterprise path
The same architecture can later support enterprise-owned hardware when governance requires control.
Proof status
Access is controlled.
TARX is in private preview. This page explains the architecture direction and acquisition use case while keeping availability and enterprise claims tied to verified proof.
Current proof
Product screenshots, local runtime flow, and clear docs for the local to hosted to enterprise-owned path.
Availability note
Runtime access is handled through controlled preview until broader availability is verified.
FAQ
Questions
answered plainly.
What does local-first AI mean?
Local-first AI means useful work starts on the user's machine by default, with hosted compute added when more headroom is needed.
Does TARX replace my app stack?
No. TARX is the runtime layer that helps an app use local execution, approved skills, memory, and hosted headroom without changing the contract.
What happens when local compute is not enough?
The app can move eligible heavy work to hosted Supercomputer headroom while keeping the runtime model understandable to the user.
Can this support enterprise deployment later?
That is the intended path. Start local, scale to hosted compute, then move onto enterprise-owned hardware when governance requires it.
Is this only for chat apps?
No. TARX is for AI work that runs, checks, remembers context, uses approved skills, and produces evidence.
Private preview
Start with
TARX runtime.
Tell us what you are building and where local-first execution should fit.