Access Claude Opus 4.8, Sonnet 4.6 and Haiku through both OpenAI-style and native /v1/messages endpoints — with prompt caching passed through — and fund it all with USDT, no card required.
Prompt caching passes straight through, so agent loops and long system prompts bill as cheap cache reads.
Call Claude via /v1/chat/completions or native /v1/messages — whichever your stack expects.
Top up in crypto, spend what you use. No card, no KYC, no regional block.
# native Anthropic Messages API curl https://halfrouter.vip/v1/messages \ -H "x-api-key: $HALFROUTER_KEY" \ -d '{"model":"claude-opus-4-8","max_tokens":1024,"messages":[...]}'
Yes. Fund your account with USDT (TRC20) and call Claude through HalfRouter — no card and no KYC required.
Yes, natively. Cache reads and writes pass through, which cuts the cost of coding agents and long, repeated system prompts.
Claude Opus 4.8, Sonnet 4.6 and Haiku, via both OpenAI-compatible and native /v1/messages endpoints.
Yes — pick the Standard tier for the biggest discount, or Dedicated for first-party upstreams with failover.
Sign up, grab a key, top up in USDT — your existing OpenAI-style code runs on GPT, Claude & Gemini at half the bill.