API Compatibility Protocols
Learn about WandAPI API endpoints, request protocols, and their relationship to model providers.
API Compatibility Protocols
WandAPI provides compatible API endpoints for different types of clients. When making a request, select the endpoint that matches the protocol used by the client.
Confirmed providers and protocols
| Provider | Anthropic Messages | OpenAI Responses | OpenAI Chat Completions |
|---|---|---|---|
| OpenAI (GPT) | × | √ | √ |
| xAI (Grok) | × | √ | √ |
| Zhipu (GLM) | √ | √ | × |
| DeepSeek | √ | √ | × |
| MiniMax | √ | √ | × |
√ means WandAPI currently has an available endpoint; × means WandAPI has not added a corresponding upstream source. It does not mean that the model itself does not support the protocol. Many Chinese models support OpenAI Chat Completions, but WandAPI has not yet added the corresponding upstream sources. When configuring a client, also check the Model Catalog and the models available to the API Key.
API endpoints
| Protocol | Request path | Description |
|---|---|---|
| OpenAI Responses API | POST /v1/responses | OpenAI Responses format |
| Anthropic Messages API | POST /v1/messages | Anthropic Messages format |
| OpenAI Chat Completions API | POST /v1/chat/completions | Legacy compatibility endpoint; corresponding Chinese-model upstream sources have not yet been added to WandAPI |
| Gemini API | POST /v1beta/models/{model}:generateContent | Google Gemini format |
“Compatible” means that requests and responses follow the corresponding protocol format. It does not mean that different models support exactly the same capabilities. For details about fields, streaming responses, and error formats, see Responses API, Messages API, and Gemini API.
Do not treat /v1/responses, /v1/messages, or /v1beta/models/... as a universal Base URL for all clients. First confirm which protocol the client uses, then enter the Base URL or complete request URL required by that client. Different clients may append different paths to the Base URL.