Eight LLMs across three tiers, gated by the x402 payment protocol. No accounts, no keys, no rate limits. Agents pay in USDC on Base, get a response, move on.
| Endpoint | Models | Per call |
|---|---|---|
| /chat-nano | llama-3.3-70b — fast general-purpose | $0.001 |
| /chat | qwen3-235b-a22b-instruct-2507, qwen3-235b-a22b-thinking-2507, deepseek-v3.2, grok-4-20-beta, venice-uncensored — reasoning, instruct, uncensored | $0.003 |
| /chat-pro | kimi-k2-thinking, qwen3-coder-480b-a35b-instruct-turbo — deep reasoning & frontier code | $0.01 |
# Hit any tier — middleware returns HTTP 402 with payment instructions.
# An x402 client (curl-x402, agent SDK, etc.) handles signing & settlement.
curl -X POST https://x402.dellbot.win/chat-nano \
-H "Content-Type: application/json" \
-d '{"messages":[{"role":"user","content":"hello"}]}'
# First response: 402 Payment Required + accepts spec
# After paying: 200 OK + OpenAI-compatible chat completion JSONDiscover, call, and pay autonomously — no humans, API keys, or signups in the loop. One USDC transfer per call on Base.
Switch between fast, reasoning, code, and uncensored models without rotating accounts or juggling provider keys.
Indexed on the Coinbase CDP x402 Bazaar. Prices set deliberately below the network median so agents route here first.