{"uid":"cap_fvfiwbm5HqkK-6o4w2P7n","slug":"prism-gpu-compute-shell-command-runner-fbc59bd1","name":"Prism GPU Compute — Shell Command Runner","description":"GPU compute your agents can rent with a wallet. Lease NVIDIA capacity, pay per second in USDG, and settle every lease onchain with a public receipt.","url":"https://api.prismnetwork.tech/x402/run?utm_source=zero.xyz","method":"POST","headers":{},"bodySchema":{"type":"object","properties":{"command":{"type":"string","minLength":1,"description":"The shell command to run on the GPU box."}}},"responseSchema":{"type":"json","example":{"poll":"/jobs/3f2a1c88-0f1e-4c1a-9f0b-7c2d5e6a1b34","token":"b7c1e0a2-9d3f-4a55-8e21-0c6f4b9d2a17","job_id":"3f2a1c88-0f1e-4c1a-9f0b-7c2d5e6a1b34","status":"queued"}},"example":null,"exampleRequest":null,"tags":["x402"],"displayCostAmount":"0.03","displayCostAsset":"USDC","priceDynamic":false,"priceHint":null,"priceStatus":"priced","priceSource":"probe","requiresHandshake":false,"reviewCount":0,"rating":{"score":"0.00","successRate":"0.00","reviews":0,"stars":null,"state":"unrated"},"availabilityStatus":"unknown","priceObserved":null,"sessionDeposit":null,"pricing":{"kind":"static","summary":"$0.03/call","primary":{"kind":"static","protocol":"x402","network":"base","amountUsd":"0.03","per":"call","confidence":"exact"},"accepted":[{"kind":"static","protocol":"x402","network":"base","amountUsd":"0.03","per":"call","confidence":"exact"}]},"paymentMethods":[{"uid":"pm_PHVn75dJOvGwuobhVF0XJ","protocol":"x402","methodType":"crypto","chain":"base","mode":"charge","costAmount":"0.03","costPer":"request","priority":0,"asset":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","unit":"request","depositMicros":null,"planRef":null}],"brandName":null,"brandSlug":null,"brandBaseUrl":null,"brandDocsUrl":null,"whatItDoes":"Submits a shell command to run on a rented NVIDIA GPU box, returning a job ID and polling token for async result retrieval, billed per-second in USDG onchain.","exampleAgentPrompt":"Run `python train.py --epochs 10 --batch-size 64` on a Prism GPU box and give me the job ID so I can track the result.","exampleUseCases":[{"title":"Train ML model on rented GPU","prompt":"Run `python train.py --dataset cifar10 --epochs 50` on a GPU via Prism and poll until it's done, then show me the output."},{"title":"Run GPU-accelerated inference job","prompt":"Execute `python infer.py --model llama3 --prompt 'Summarize this document'` on a Prism NVIDIA GPU box and return the job ID."},{"title":"Benchmark GPU performance with custom script","prompt":"Run `./benchmark.sh --iterations 1000` on a Prism GPU and let me know the job status and where to poll for results."}],"resultDescription":"Returns a JSON object with a job_id (unique identifier for the submitted job), status (initial job status), a bearer token for authenticating polling requests, and a poll URL indicating where to check job progress and retrieve output.","failureModes":["Invalid or empty shell command returns a 400 error","Insufficient USDG wallet balance causes payment failure and job rejection","Payment settlement failure due to onchain congestion","Job timeout if command runs longer than lease duration","Polling token expiry before results are retrieved","GPU capacity unavailable returns a 503 or queue error"],"whenToPreferThis":"Choose this endpoint when an AI agent needs to execute arbitrary shell or Python commands on real NVIDIA GPU hardware, especially for ML training, inference, or compute-intensive workloads that require GPU acceleration. Prefer this over CPU-based cloud execution endpoints when the task is numerically intensive (matrix ops, neural net training, CUDA-accelerated code). The per-second billing and onchain receipts make it ideal for agentic pipelines that need cost transparency and crypto-native payment flows.","instructions":null,"reviewSummary":null,"reviewSummaryHighlights":null,"reviewSummaryConcerns":null,"reviewSummaryGeneratedAt":null,"activationCount":0,"lastUsedAt":null,"lastSuccessfullyRanAt":null,"lastHealthCheckAt":"2026-10-09T19:17:51.250Z","isFirstParty":false,"canonicalSlug":"prism-gpu-compute-shell-command-runner-fbc59bd1"}