{"service":"ComputeIndex","version":"1.0.0","tagline":"Regional cost of powering AI compute, agent-callable.","description":"The regional cost of powering AI compute, made agent-callable. One x402 call returns the energy cost to run a GPU for one hour (H100, H200, B200, A100) in each US wholesale power region (CAISO SP15/NP15; ERCOT Houston, North, South, West, hub average), each region's basis versus the average, the cheapest and priciest region, the spread translated to a 1,000-GPU cluster per month, live GPU spot rental prices (Vast.ai, RunPod, AWS spot, Lambda reference) and grid stress (EIA-930). Assumptions are published in every response (GPU board kW x PUE x real-time $/MWh). Compute is becoming an economic resource with regional pricing; this is the price-discovery input a data-center siting, workload-scheduling, compute-procurement or compute-finance agent needs. $0.10 in USDC on Base or Solana via x402. Built on LSO ComputePulse + SparkPulse + GridPulse.","when_to_call":["Find the cheapest US power region to run GPUs right now","Price the regional energy basis for a compute cluster or GPU-backed financing","Compare electricity cost against GPU spot rental prices"],"price":"$0.10 USDC via x402 (Base or Solana)","payment":{"protocol":"x402","network":"base-mainnet","asset":"USDC","payTo":"0x52Ab53912D37759B2ad364f22dD06B16714b6C06"},"endpoints":{"/index":"energy cost per GPU-hour by US power region + regional basis (no params)  $0.10","/health":"status (free)"},"provider":"LoneStarOracle","catalog":"https://lonestaroracle.xyz","a2a":"https://lonestaroracle.xyz/.well-known/agent-card.json"}