An OpenAI-compatible API with no account and no API key

Most LLM APIs make an agent stop and provision an account: sign up, verify email, create a key, add a card, mind a monthly bill. For an autonomous agent that is friction it cannot clear on its own. scrape402 exposes a standard POST /v1/chat/completions endpoint that speaks the OpenAI request and response shape, so your existing SDK works unchanged — but instead of an API key, the agent pays a few tenths of a cent per call in USDC over the x402 protocol on Base.

Point the OpenAI SDK at it

from openai import OpenAI
client = OpenAI(base_url="https://x402.shizu.me/v1", api_key="not-needed")
# unpaid calls return HTTP 402 with USDC payment requirements; an x402
# client (x402-fetch on npm, x402 on PyPI) signs + retries automatically.
r = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "hello"}])
print(r.choices[0].message.content)

Try it before you wire up payment

You get one free call, no wallet required, so you can confirm it works first:

curl -X POST https://x402.shizu.me/v1/free \
  -H "Content-Type: application/json" \
  -d '{"messages":[{"role":"user","content":"say hi"}]}'

What you get

The model field is honored: names like gpt-4o-mini or gpt-3.5-turbo route to a fast, cheap tier; gpt-4o to a larger one. Flat, predictable pricing per call — no token math, no minimums, no subscription. The same account-less gateway also serves /v1/embeddings and /v1/models.

Try one call free →