{ "page": "https://x402-list.com/apis/vps-api", "title": "VPS and sandbox APIs you can pay per call", "answer": "As of October 26, 2026, x402-list compares 5 VPS and sandbox APIs with documented compute routes. The median selected quote is $0.018. Code sent to a sandbox with each call, a command in a running sandbox and a job on a rented GPU are different products: compare time limits, network access, what survives between calls and billing on failure.", "date_modified": "2026-10-26", "license": "https://creativecommons.org/licenses/by/4.0/", "cite_as": "x402-list, \"VPS and sandbox APIs you can pay per call\", https://x402-list.com/apis/vps-api, data as of 2026-10-26. Licensed CC BY 4.0.", "how_to_cite": "Quote the number with its date and link the page URL. Data is CC BY 4.0: attribution to x402-list is required.", "methodology": "https://x402-list.com/methodology", "criteria": { "listed": "online, payment-ready, 30-day uptime >= 90%, known price, description >= 80 characters", "price": "cheapest current compute route of each service, USD per call (not its cheapest tool of any kind); services without one are left out", "generalist": "a gateway matching more than three unrelated capabilities: listed, never called the cheapest", "buyers_30d": "distinct wallets that paid the service on-chain in the last 30 days (measured floor)", "best": "most distinct buyers in 30 days; ties go to higher 30-day uptime" }, "in_index": true, "stats": { "services": 5, "with_buyers_30d": 3, "median_price_usd": 0.018, "min_price_usd": 0.001, "max_price_usd": 0.4, "generalists": 2, "left_out_no_task_endpoint": 6 }, "faq": [ { "question": "Is there a free VPS or sandbox API in this comparison?", "answer": "The 5 listed services have positive per-call prices for their task endpoints. A free test network, a free retry or an unbilled failure is a provider statement and does not make the paid route free. Check the terms for the exact request." }, { "question": "How much do 100,000 sandbox or GPU calls cost?", "answer": "Using the median selected route, 100,000 calls total $1,800.00 at $0.018 per call. That assumes one call per job; GPU work longer than one block, retries and calls that only check a result change the total." }, { "question": "Is a sandbox API the same as a VPS?", "answer": "No. A VPS is a server you rent for a period and manage yourself, with a disk and processes that persist. A sandbox route runs one job per call and returns its output. Ask what persists between calls before you build on either." }, { "question": "Can code in these sandboxes reach the internet?", "answer": "Only where the provider says so. A route description can state that the sandbox has no network access; the guide above names the services that say so. Where nothing is stated, ask before you send code that installs packages or calls an API." }, { "question": "Do files persist between calls?", "answer": "Do not assume either way unless the route description says so, and that holds for a sandbox that is already running too. Ask the provider, save anything you need to resume elsewhere, and send it back in with the next call." }, { "question": "How long can one job run?", "answer": "Check the route description and the service notes on this page: a route can state a time cap, sell a fixed GPU block or state no limit at all. Where no limit is stated, ask the provider, and split long work into calls that each finish within the limit." }, { "question": "Can I rent a GPU per call?", "answer": "As of October 26, 2026, 2 listed services document a priced GPU route. Check the GPU model, the length of the block and how the workload and its results are moved before you pay." }, { "question": "Is it safe to run untrusted code in these sandboxes?", "answer": "Isolation is a provider statement; x402-list has not tested it. Keep secrets out of the code, use short-lived credentials when a job must call out, and treat the output as untrusted input to your own systems." }, { "question": "Does an HTTP 402 check prove my code will run?", "answer": "It shows that the route asks for payment. It does not show that a settled request runs your code correctly. The listing reports monitoring and payment observations separately from provider claims." }, { "question": "Does latency show how long a job takes?", "answer": "No. The latency column comes from unpaid monitoring checks. A paid job also waits for the sandbox or the GPU to start and for your code to finish, which depends on the work you send." } ], "services": [ { "name": "vaaya", "reseller_of_brand_via": null, "url": "https://x402-list.com/services/vaaya", "price_usd_per_call": 0.01, "networks": [ "Base" ], "uptime_30d": 100, "latency_ms": 451, "buyers_30d": 57, "last_checked_at": "2026-10-26T19:03:27.362Z", "what_it_does": "Provider lists a route that runs a command in an already running Modal sandbox, beside other resold tools.", "generalist": true, "price_endpoint": { "method": "POST", "path": "/api/run/modal/sandbox-exec" }, "task_endpoints": [ { "method": "POST", "path": "/api/run/modal/sandbox-exec", "price_usd_per_call": 0.01 } ], "endpoints_listed": 7 }, { "name": "scrape402", "reseller_of_brand_via": null, "url": "https://x402-list.com/services/scrape402", "price_usd_per_call": 0.018, "networks": [ "Base" ], "uptime_30d": 100, "latency_ms": 109, "buyers_30d": 26, "last_checked_at": "2026-10-26T19:03:14.890Z", "what_it_does": "Provider lists a code sandbox for Python, JavaScript, Bash, R or Java with a 25-second cap, beside scraping and market data.", "generalist": true, "price_endpoint": { "method": "POST", "path": "/sandbox" }, "task_endpoints": [ { "method": "POST", "path": "/sandbox", "price_usd_per_call": 0.018 } ], "endpoints_listed": 40 }, { "name": "Kilawatt Cloud", "reseller_of_brand_via": null, "url": "https://x402-list.com/services/kilawatt-cloud", "price_usd_per_call": 0.4, "networks": [ "Base" ], "uptime_30d": 98.7, "latency_ms": 6459, "buyers_30d": 6, "last_checked_at": "2026-10-26T19:10:12.874Z", "what_it_does": "Provider lists one GPU compute route, described as one NVIDIA H100 for 600 seconds, and says it routes work across several GPU clouds.", "generalist": false, "price_endpoint": { "method": "POST", "path": "/api/public/x402/exec" }, "task_endpoints": [ { "method": "POST", "path": "/api/public/x402/exec", "price_usd_per_call": 0.4 } ], "endpoints_listed": 1 }, { "name": "Prism Network", "reseller_of_brand_via": null, "url": "https://x402-list.com/services/prism-network", "price_usd_per_call": 0.03, "networks": [ "Base" ], "uptime_30d": 100, "latency_ms": 222, "buyers_30d": 0, "last_checked_at": "2026-10-26T18:54:17.373Z", "what_it_does": "Provider lists one shell command on a rented GPU, charged only on success, beside two model-inference routes.", "generalist": false, "price_endpoint": { "method": "GET", "path": "/x402/run" }, "task_endpoints": [ { "method": "GET", "path": "/inference/v1/batch", "price_usd_per_call": 0.012144 }, { "method": "GET", "path": "/inference/v1/inference", "price_usd_per_call": 0.012144 }, { "method": "GET", "path": "/x402/run", "price_usd_per_call": 0.03 } ], "endpoints_listed": 3 }, { "name": "agent-exec", "reseller_of_brand_via": null, "url": "https://x402-list.com/services/agent-exec", "price_usd_per_call": 0.001, "networks": [ "Base", "SOL" ], "uptime_30d": 99.9, "latency_ms": 410, "buyers_30d": 0, "last_checked_at": "2026-10-26T18:54:26.544Z", "what_it_does": "Provider lists Python or Node code execution in an isolated sandbox without network access, beside storage and directory tools.", "generalist": false, "price_endpoint": { "method": "POST", "path": "/execute" }, "task_endpoints": [ { "method": "POST", "path": "/execute", "price_usd_per_call": 0.001 } ], "endpoints_listed": 15 } ] }