Same drop-in idea — one base URL, every major model — but you top up in USDT instead of a card, get native Claude prompt caching, and choose price vs. stability per request across three tiers.
OpenAI-style chat and native /v1/messages. Point your base URL — keep your SDK and code.
Top up in USDT. No credit card, no identity forms, no regional lockout.
Standard for the cheapest route, Dedicated for pure first-party upstreams with failover.
# pip install openai — change one line client = OpenAI(base_url="https://halfrouter.vip/v1", api_key="YOUR_HALFROUTER_KEY") client.chat.completions.create(model="claude-sonnet-4.6", messages=[...])
Yes. It exposes standard /v1/chat/completions and native Anthropic /v1/messages. Change base_url to https://halfrouter.vip/v1 and keep your existing OpenAI SDK or tools.
You pay in crypto (USDT) with no card and no KYC, Claude prompt caching passes through natively, and you pick a price/stability tier per request instead of a single fixed markup.
GPT-5.5 / 5.4 / codex, Claude Opus 4.8, Sonnet 4.6 and Haiku, Gemini 3 Pro & Flash, plus DeepSeek, Qwen, Kimi and GLM.
No. It is a one-line base_url change. Works with the OpenAI SDKs, Cline, Roo, Aider, Cursor, LibreChat and anything that speaks the OpenAI API.
Sign up, grab a key, top up in USDT — your existing OpenAI-style code runs on GPT, Claude & Gemini at half the bill.