CapIX is the OS for decentralised compute. Live decentralised (and quantum) nodes are active now; CapIX OS finds spare room, recycles idle capacity, and routes compute and inference by live price — billed per minute in SOL, USDC, or CPX. The OS ships three apps: Inference, CapIX Cloud, and AI Labs.

Anyone who needs machine time can deploy on indexed capacity; GPU hosts, CPU hosts, labs, endpoint operators, and home providers can provide spare supply after verification.

Smart Route is CapIX's proprietary job-routing system. It evaluates task shape, price, reliability, latency, and available capacity to produce a receipt-backed delivery path for compute or inference.

UsePod adds another live inference market to Smart Route

More supply, one Capix endpoint

UsePod is being added as another OpenAI-compatible marketplace lane behind Capix Smart Route. Capix discovers eligible models, compares the lane with other available routes and can fail over without asking a developer to rewrite their application.

The upstream identity and credential remain inside the control plane. Customers see a Capix route, a Capix charge and an anonymised route label rather than a provider-specific secret.

A path for people who already serve models

Operators running vLLM, llama.cpp, LM Studio or Ollama can follow the UsePod Host onboarding from the Capix provider page. UsePod manages its enrollment, provider bond, marketplace pricing and USDC payout; Capix keeps that external balance separate from direct Capix provider earnings.

The demand lane is protected by a production feature gate, a server-side proxy token and hard input/output price ceilings. It cannot silently become an unlimited-spend route.

Sell your inference →