◎
One interface
Keep the client you already use. Models and live providers meet behind one familiar API.
● ABOUT TOKEN2TOKEN
Compute should move as freely as ideas.
Token2Token is TokenTaiwan's shared inference network: one compatible API that connects official infrastructure with community GPUs around the world.
⌂ Taiwan-built⌘ OpenAI compatible◇ Accountable ledger
⌁
Reciprocal compute
Put an idle GPU to work, earn Indigo for verified output, and use it across community routes.
◇
Explicit trust
Official, managed, and community endpoints remain clearly labeled so teams can route with intent.
TOKEN2TOKEN · TAIWAN
Make more of the world's GPUs useful.
Taiwan sits at the center of the hardware that powers AI. We are building the connective layer that lets developers reach that compute, lets owners contribute capacity, and keeps price, location, and trust visible for every route.
GPUshared supply
→T2Tone router
→APIlive demand
✦ OUR PRINCIPLES
Open by design. Clear by default.
A useful compute network needs more than available hardware. It needs understandable incentives and honest boundaries.
Compatible, not captive
Open clients and familiar APIs make it easy to join or leave without rebuilding your stack.
Local when it matters
Taiwan Official routes provide a deliberate option for residency and contractual protection.
Trust must be visible
Provider tier, location, verification, pricing, and privacy claims stay attached to every endpoint.
Earned value stays useful
Indigo turns verified community inference into service credit that can fund more community inference.
◇
KNOW THE BOUNDARY
Read the routing guide →Community does not automatically mean confidential.
Ordinary community nodes may observe plaintext inference data. Sensitive workloads belong on contract-protected Official routes or verified confidential hardware.
TOKEN2TOKEN BETA
Bring supply and demand onto one network.
Build through one API or connect an idle GPU and start contributing.