A drop-in gateway for indie devs and R&D teams. Point your base URL here and keep your existing OpenAI SDK — native Claude prompt caching, three reliability tiers, and you pay in crypto: no card, no KYC.
If it speaks the OpenAI API, it speaks HalfRouter. Copy, paste, ship.
# Chat completions — OpenAI-compatible
curl https://halfrouter.vip/v1/chat/completions \
-H "Authorization: Bearer $HALFROUTER_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "claude-sonnet-4.6",
"messages": [{"role":"user","content":"Hello!"}]
}'# pip install openai — no other changes
from openai import OpenAI
client = OpenAI(
base_url="https://halfrouter.vip/v1",
api_key="YOUR_HALFROUTER_KEY",
)
resp = client.chat.completions.create(
model="gpt-5.5", # or claude-opus-4-8, gemini-3-pro-preview…
messages=[{"role":"user","content":"Hello!"}],
)
print(resp.choices[0].message.content)// npm i openai — drop-in
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://halfrouter.vip/v1",
apiKey: process.env.HALFROUTER_KEY,
});
const r = await client.chat.completions.create({
model: "claude-sonnet-4.6",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(r.choices[0].message.content);# Cline / Roo Code → Settings → API Provider
API Provider : OpenAI Compatible
Base URL : https://halfrouter.vip/v1
API Key : YOUR_HALFROUTER_KEY
Model ID : claude-sonnet-4.6 # great for agentic coding + caching# Aider — env vars, then run
export OPENAI_API_BASE="https://halfrouter.vip/v1"
export OPENAI_API_KEY="YOUR_HALFROUTER_KEY"
aider --model openai/claude-sonnet-4.6# Cursor → Settings → Models → OpenAI API Key → Override base URL
Base URL : https://halfrouter.vip/v1
API Key : YOUR_HALFROUTER_KEY
# add a custom model name, e.g. gpt-5.5 or claude-opus-4-8OpenAI-compatible endpoints for chat and native /v1/messages. Change one base URL — no SDK, no rewrites.
Prompt caching passes straight through. Agent loops and long system prompts get cache reads — not full re-bills.
Top up in USDT. No US card, no identity forms, no regional lockout. Pay-as-you-go, spend what you use.
Pick Standard for the cheapest route or Dedicated for pure first-party upstreams with cascading failover.
Same model, your choice of price vs. stability. Billed in USD against list price — pick per request via the tier group.
Coding agents resend the same system prompt and file context every turn. With native Claude caching, those tokens come back as cheap cache reads instead of full-price input — the difference compounds over a long session.
Sign up, grab a key, top up in USDT, and your existing code is live on GPT, Claude and Gemini — at half the bill.