LLM API — pay per request

One OpenAI-compatible endpoint. No signup, no API key, no KYC. Send a request, get HTTP 402 with payment terms, pay, retry with a payment header.

💵 USDC via x402⚡ Nano (XNO)🔒 No KYC🧠 13 models

USDC — x402

$0.005 per request
Base · Polygon · Arbitrum
Machine-native: the 402 body carries the terms

Nano — XNO

0.001 XNO per request
Fee-free and instant
Retry with the send block hash

Flat price per call. Long prompts and long completions cost the same — there is no token metering.

autorouter picks the best available
gemini-3.6-flashGoogle, fast
gemini-3.5-flashGoogle, previous
gpt-oss-120bOpen weights, 120B
deepseek-v4-flashfast general
deepseek-v4-prostronger reasoning
deepseek-r1chain-of-thought
kimi-k2.6long context
kimi-k2.7-codecoding
qwen3-coder-480bcoding, 480B
glm-4.5general
llama-3.1-8bsmall and cheap
minimax-m3general

How it works

1. Send your request to POST /v1/chat/completions with no payment header. The endpoint replies 402 with the price, the accepted schemes and the payTo address.
2. Pay 0.005 USDC (x402, on any listed chain) or 0.001 XNO (Nano).
3. Repeat the identical request with the payment header — PAYMENT-SIGNATURE for x402, X-Nano-Payment: <send block hash> for Nano. You get the completion back.
curl -X POST https://pyfile-agent.taile3ff35.ts.net/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"auto","messages":[{"role":"user","content":"hello"}]}'

# → 402 with payment terms; pay; then repeat with the payment header
Client libraries. Wallets and x402-capable clients handle the 402 handshake automatically. The endpoint is a drop-in for OpenAI SDKs — point the base URL here, keep the body identical.

FAQ

Is there an API key or signup?
No. There is no account, no key and no email. You send a request, get HTTP 402 with payment instructions, pay, and retry the same request with a payment header.
How much does one request cost?
One flat price per request regardless of prompt or completion length: 0.005 USDC via x402, or 0.001 XNO via Nano. Long prompts do not cost more.
Which chains and tokens can I pay with?
USDC on Base (eip155:8453), Polygon (eip155:137) or Arbitrum (eip155:42161) through the x402 protocol, or XNO on the Nano network. The 402 response lists every accepted option with the exact payTo address.
Is this OpenAI-compatible?
Yes. POST /v1/chat/completions takes the same JSON body as the OpenAI Chat Completions API, so existing SDKs work once the base URL is pointed here and the payment header is attached.
What happens if I send a request without paying?
You get HTTP 402 and a JSON body describing the price, the accepted payment schemes and the payTo address. Nothing is charged until a valid payment header is presented.

Machines can discover the terms without any documentation: GET / returns the service descriptor, GET /v1/models the model list, and any unpaid POST returns the full 402 payment requirements.
This page is part of the Web Dev Toolkit.