LLM inference your agent
can buy one request at a time.
An OpenAI-compatible chat completions endpoint billed per call in stablecoins over the x402 protocol. No account, no dashboard, no API key to provision — the payment is the authorization.
Just call it 🪼
Why no signup
Every hosted LLM API assumes a human signs up first. An autonomous agent that needs one completion cannot create an account, confirm an email, enter a card and provision a key. That signup flow is a structural barrier, not a technical one. LLM2API removes it.
Quickstart
If your code already calls OpenAI, change the base URL.
# Call once, paying with x402 (Bankr CLI) bankr x402 call \ https://x402.bankr.bot/0x444fadbd6e1fed0cfbf7613b6c9f91b9021eecbd/llm2api/v1/chat/completions \ -X POST \ -d '{"messages":[{"role":"user","content":"Summarise this in one sentence."}]}'
Pricing
Flat per request. No subscription, no minimum, no expiry.
Bankr x402 — USDC (Base)
The default rail. Recommended for agents and MCP workflows.
JPYC x402 — Polygon
For yen-denominated settlement.
Conventional key-based access
If you would rather use a normal API key and monthly plan.
API
POST {base}/v1/chat/completions
| Field | Type | Required | Notes |
|---|---|---|---|
messages | array | Yes | {role, content} objects. Roles: system, user, assistant. |
stream | boolean | No | SSE streaming. Default false. |
temperature | number | No | 0.0–2.0. Default 0.7. |
max_tokens | integer | No | A hard cap of 2,048 applies regardless. |
- Input — 4,000 characters total across all messages, 20 messages maximum
- Output — hard cap of 2,048 tokens per request
- Model — every paid rail is served by
deepseek-v4-flash
These caps are deliberate. They keep a single request from monopolising the backend and keep the flat per-call price honest.
Honest limitations
- This is a small, self-operated service. It is not a hyperscaler and makes no uptime guarantee.
- The input and output caps above are hard limits. Long-document workloads are not a fit.
- Requests are proxied to an upstream provider. Do not send data you would not send to a third-party LLM API.
The growing Kurage series
LLM2API is the inference layer of the Kurage ecosystem. The same plumbing powers the products below.
URLAI is the ecosystem token
URLAI is the token for spreading the Kurage ecosystem. Beyond URL2Earn, it is also distributed to kfreqai ambassadors.
- Ambassadors trade crypto and FX with kfreqai and share their results.
- That improves kfreqai's profitability and spreads awareness, aiming at a project that eventually earns for many people.
- URLAI aims to be part of the ecosystem's tokenomics, distributed for that spreading and contribution.
* URLAI is a token of the Kurage ecosystem as a whole, not tied to any single product. See 🧭 Vision — why we share (JA) / 📐 Design — the tokenomics in numbers (JA). Its price and liquidity fluctuate with the market and no monetary value is guaranteed. Claiming is not investment advice.