Cortex

Every prompt.One cluster.

Stake SOL into model clusters. Mine inference credits and burn them for real model calls. Audit any token with the Trench gateway.