███╗ ██╗███████╗██╗ ██╗██╗ ██╗███████╗ ██╗ ██╗██╗ ██╗██████╗
████╗ ██║██╔════╝╚██╗██╔╝██║ ██║██╔════╝ ██║ ██║██║ ██║██╔══██╗
██╔██╗ ██║█████╗ ╚███╔╝ ██║ ██║███████╗ ███████║██║ ██║██████╔╝
██║╚██╗██║██╔══╝ ██╔██╗ ██║ ██║╚════██║ ██╔══██║██║ ██║██╔══██╗
██║ ╚████║███████╗██╔╝ ██╗╚██████╔╝███████║ ██║ ██║╚██████╔╝██████╔╝
╚═╝ ╚═══╝╚══════╝╚═╝ ╚═╝ ╚═════╝ ╚══════╝ ╚═╝ ╚═╝ ╚═════╝ ╚═════╝
▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
█ UNIFIED AI INFERENCE │ INTELLIGENT MODEL ROUTING █
█ 42 MODELS AVAILABLE │ AUTO MODEL SELECTION █
█ STREAMING SUPPORT │ OPENAI COMPATIBLE API █
▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀
> INTELLIGENT ROUTING
Auto-selects the best model based on your request. Code → Qwen Coder, Math → DeepSeek R1, General → GLM
> MULTI-PROVIDER
Access DeepSeek, Qwen, Llama, NVIDIA Nemotron, Google Gemma, Kimi, GLM through one API
> OPENAI COMPATIBLE
Drop-in replacement for OpenAI API. Same request/response format, instant migration
> STREAMING SSE
Real-time token delivery via Server-Sent Events. Support for reasoning tokens
$ curl -X POST /v1/chat/completions \
-d '{"messages": [{"role": "user", "content": "Hello!"}]}'
{
"router": {"mode": "AUTO", "selected_model": "glm-4.5"},
"choices": [{"message": {"content": "Hello! How can I help?"}}]
}