All the frontier models, one key.
DeepSeek, Kimi, GLM, Qwen, Doubao, Kling + Claude, GPT-5, Gemini — 15+ models, one OpenAI-compatible API. Drop into Cursor and Claude Code with one BASE_URL change. Top up with WeChat / Alipay / USDT / Stripe — no card required.
export OPENAI_BASE_URL=https://routify.bytedance.city/v1
export OPENAI_API_KEY=rtf_...
# Drop-in: every OpenAI SDK now hits RoutifyDrop-in for Cursor · Claude Code · Cline · Aider
Why Routify
We obsess over three things: coverage, transparent pricing, working payments.
Every frontier model, one key
7 Chinese: DeepSeek, Kimi, GLM, Qwen, Doubao, MiniMax, Kling + 3 global: Claude, GPT, Gemini. No juggling accounts, no juggling BASE_URLs. New model versions ship within 24h.
Drop into Cursor & Claude Code
Cursor: change OPENAI_BASE_URL. Claude Code: change ANTHROPIC_BASE_URL. No VPN, no rate limits, no ban risk. Smart routing fails over to a healthy backup automatically.
Honest pricing, working payments
Global models (Claude / GPT / Gemini) pass through at official rates. Chinese models marked up by region: 0-15% for CN users, 30-80% for global. See /pricing for the per-model breakdown. Zero topup fees.
Models
15+ frontier models, one API
Smart routing picks the lowest-cost healthy channel automatically. New versions land within 24h.
Pricing
Pay-as-you-go, per token.
Buy credits, spend them on any model. Zero topup fee, credits never expire.
DeepSeek V3.2
Cheap, fast, GPT-4-class for code & chat. Best for high-volume.
- 65k context
- Streaming · tools · JSON mode
- +30-80% markup overseas (still 13× cheaper than GPT-5)
Claude Opus 4.7
$15/MPremium reasoning at official Anthropic rates. Zero markup.
- 200k context
- Anthropic official rate, pass-through
- EU / US / Asia endpoints