Turn idle GPU power into useful AI compute. Explore how inference, verification, and independent providers fit together.
Three ways into the network
One network, designed for everyone who creates, supplies, or owns the intelligence running through it.
FOR DEVELOPERS
The core idea: route open-model requests to independent GPUs, match the workload, and verify the result. Explore the path from a request to useful compute.
Start buildingDiagram: an API request is routed across eight GPU nodes and recombined into one verified response.
FOR GPU OWNERS
Choose a browser or machine setup. Work must be requested, completed, and verified before any provider reward can be earned.
Start earningDiagram: three chosen workload levels converge into a verification checkpoint. These are mechanism inputs, not live utilization.
FOR NETWORK OWNERS
Explore the proposed revenue rule: half of network margin for token supply reduction, half for stakers. Change the input to understand the split.
Explore the mechanismDiagram: network revenue splits into exactly two halves — 50% distributed to stakers, 50% of the token burned from supply.
Build on it. Power it. Own a piece of it.
Enter OpenweftView the network