▸ __BREADCRUMB__ usage
new: genuinely free AI models are live — try surp/free + see live budgets · token-gating prototype · vote on SRP

pay-per-request LLM API

pay as you go. literally.

Most LLM APIs are subscriptions: a monthly bill whether you use it or not. surp.ivc.lol is a pay-per-request LLM API — you pay only for the requests you make, in USDC on Base, via x402. No subscription, no credit card, no account, no API key.

how pay-per-request pricing works here

Every request is priced at the live market rate of the model that serves it. You choose a combo (e.g. surp/best-coding), and the gateway routes to the cheapest model in that class at that second. The price is quoted in the 402 response, you sign the USDC transfer, and the model answers.

~1¢
per request
0
monthly fees
0
accounts
150+
models, dynamically priced

when pay-per-request makes sense

  • Agents and bots — autonomous workloads that shouldn't carry a subscription.
  • Sporadic use — you call an LLM occasionally; a monthly plan is wasted money.
  • Cost-sensitive products — every cent counts; pay only for real usage.
  • Experiments and demos — spin up, try, tear down, no billing baggage.

no wallet? use a prepaid key

If you'd rather not hold crypto, we also offer prepaid API keys — top up a balance and use a standard Authorization: Bearer header. Same per-request pricing, same dynamic routing. Create a key »

try it

curl -X POST https://surp.ivc.lol/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"surp/best-chat","messages":[{"role":"user","content":"hi"}],"max_tokens":50}'

First response is a 402 with the price. Sign with an x402 client and the completion streams back. See connect your agent for full setup.

related: what is x402? · x402 LLM API · x402 gateway · cheapest LLM API · docs