About SilvaMux
SilvaMux is an enterprise-grade LLM gateway: one unified API gives you access to multiple model providers, so your business can use chat, image generation, video generation, and 3D generation without integrating each vendor separately.
Why SilvaMux
- Unified access: one API key and one base URL, compatible with mainstream protocols such as OpenAI, Anthropic, and Volcengine Ark. Existing SDK code migrates by changing only the base URL and key.
- Model aggregation: aggregates mainstream models such as MiniMax, Kimi, GLM, DeepSeek, Qwen, and Doubao, invoked uniformly by call name, with automatic routing to available lines.
- Standards compatible: chat APIs are compatible with OpenAI Chat Completions, Anthropic Messages, Volcengine v3, and OpenAI Responses formats; image and video APIs follow the OpenAI Images style.
- Enterprise reliability: millisecond-level response, highly available architecture, usage-based billing with organization-level discounts, bill exports, and invoices.
- Officially authorized: all models are accessed through official or legitimate channels.
Core Concepts
| Concept | Purpose |
|---|---|
| Workspace | The owning unit of accounts, balance, bills, and identity verification |
| Project | Isolates businesses or environments; usage is attributed per project |
| API Key | Used for model calls, bound to one project |
| AK/SK | Used by compatible APIs that require Access Key ID and Secret Key signing |
| Model call name | The name used in the request body model field |
| API format | Determines the request path, request body, and response structure |
Order of decisions: pick a model → copy its call name → confirm supported formats → use the matching API.
Next Steps
- First time here? Start with the Quick Start.
- Want to understand costs? See Billing Overview.
- Calls failing? See Error Codes.