Providers

Provider Architecture

Each provider in Yunxin has an independent adapter that interfaces with the provider's native API. This ensures:

100% API fidelity — All provider-specific features are preserved
Optimal format selection — Yunxin uses each provider's best-suited API format, not blindly defaulting to OpenAI compatibility
Independent updates — New provider features can be supported without affecting other providers
Graceful degradation — If a provider is down, the system can fall back to alternatives

All providers are equal with full API coverage. Models are synced automatically — use the Models API to query available models.

Provider	Key Features
Alibaba (Tongyi)	Multilingual, reasoning mode
Anthropic	Extended Thinking, long context, Computer Use
Azure AI Foundry	Enterprise security, multi-provider, regional deployment
Cloudflare	Edge inference, global CDN
DeepSeek	Open-source reasoning, cost-effective, strong math/coding
GitHub Models	Free tier, GitHub integration, developer-friendly
Google (Gemini)	Long context, native multimodal, Deep Thinking
Groq	LPU-accelerated, ultra-fast inference
Hugging Face	Largest model hub, Serverless Inference API
LM Studio	Desktop GUI, local inference, model management
MiniMax	Multimodal generation, TTS, video generation
Mistral AI	Open-source, code generation, real-time audio
ModelScope	Chinese models, multimodal, free tier
Moonshot	Long context, document analysis, strong reasoning
NVIDIA NIM	GPU-accelerated, enterprise-grade
Ollama	Simple local deployment, CLI tool, model library
OpenAI	Full suite, Responses API, Realtime API, vision, embeddings
OpenRouter	Model marketplace, price comparison, automatic fallback
vLLM	High-throughput serving, PagedAttention, production deployment
Volcengine	ByteDance ecosystem, video generation, simultaneous interpretation
xAI	Deep Thinking, long context, real-time knowledge, structured output
Z.AI	Professional-grade models, enterprise-focused, high reliability
Zhipu AI	Chinese-optimized, strong bilingual support

All providers support OpenAI-compatible API format through Yunxin, ensuring seamless integration with your existing code.

Providers are configured at the system level by administrators:

Check provider status via the API:

curl https://api.yuhuanstudio.com/v1/providers \
  -H "Authorization: Bearer YOUR_API_KEY"

Response includes real-time health status for each provider.