CYRUS//RELAY

OpenAI-compatible gateway into multi-provider LLM capacity.
One key, one endpoint, crypto-only billing. No email. No KYC. No logs beyond usage accounting.

Endpoint

resolving live endpoint…

Use it — one-line swap from the official SDK

from openai import OpenAI

client = OpenAI(
    base_url="…/v1",
    api_key="sk-cyrus-…",   # get a key below
)

r = client.chat.completions.create(
    model="gemini-3.6-flash",     # free tier
    messages=[{"role":"user","content":"hello"}],
)
print(r.choices[0].message.content)

Models & pricing

Get a key

Free trial: one click → instant key, all free-tier models, no payment, no signup.

…

Free tier: gemini-*-flash + openrouter :free models · 20 req/min · no card · no email.

Paid credits: USDC/USDT on Base, Arbitrum, Polygon, BSC or Ethereum. $2 minimum. Exact-amount matching — your key appears on the order status page ~30s after your transfer confirms.

Help / support: @hetsvdbot on Telegram. Repo: ghassan-gaidi/cyrus-relay

Try it live

Real completion through our endpoint right now — no signup, no key, no anything:

click RUN →

anonymous tier · 10 req/min per IP · free-tier model · or get a key for 20 RPM, no daily reset, and premium models

Why pay

• Devs in unsupported regions / without cards need inference too — settlement is on-chain, identity-free.
• Free tier is rate-limited to 20 req/min and flash-class only. Paid credits unlock pro models (GPT/Claude/Gemini-Pro class), no rate ceiling, priority routing.
• Credits never expire. Abuse kills the key without refund.
• Founder pricing: first 50 paid keys keep these rates for life.