The complete /v1 reference.

Endpoints

Endpoint	Method	Purpose
/v1/models	GET	List available models
/v1/chat/completions	POST	Chat completions (text, streaming)
/v1/embeddings	POST	Embedding vectors
/v1/completions	POST	Legacy completions

Authentication

Send your API key as a Bearer token:

Authorization: Bearer sk-live-...

Keys are created in the dashboard. Never expose keys in client-side code; proxy through your backend.

POST /v1/chat/completions

Request body (OpenAI-compatible):

{
  "model": "plugsky-pro",
  "messages": [{"role": "user", "content": "Hello"}],
  "max_tokens": 200,
  "temperature": 0.7,
  "stream": false
}

Response includes id, object, created, model (echoes the upstream model), choices[] with message.content, and usage token counts.

Streaming

Set "stream": true. Responses are SSE chunks:

data: {"id":"chat-...","object":"chat.completion.chunk","choices":[{"delta":{"content":"..."}}]}

POST /v1/embeddings

{
  "model": "plugsky-embed-multilingual",
  "input": ["hello", "مرحبا"]
}

Returns data[].embedding vectors — 2,048 dims for plugsky-embed, 4,096 for plugsky-embed-nim and plugsky-embed-multilingual.

Model list

30+ models: plugsky-micro, plugsky-lite, plugsky-plus, plugsky-pro, plugsky-max, plugsky-frontier, plugsky-reasoning, plugsky-kimi, plugsky-deepseek-pro/flash, plugsky-gpt-oss, plugsky-qwen-next, plugsky-coder, plugsky-minimax, plugsky-vision-fast, plugsky-llama4, plugsky-longctx, plugsky-mistral-medium/small, plugsky-gemma-4, plugsky-nano/tiny/phi, plugsky-ultra, plugsky-thinker, plugsky-coder-ultra, and more. The full catalog with specs is on the models page.

Errors

Errors use OpenAI-style JSON:

{"error":{"message":"Invalid API key format. Keys start with `sk-live-`.",
  "type":"authentication_error","code":"401"}}

Common codes: 401 authentication, 404 unknown model, 429 rate limit, 500 upstream failure (auto-retry is safe).

Reference FAQ

Is the API compatible with LangChain?

Yes — LangChain's OpenAI integration accepts base_url. See the chat completions guide.

Do you support vision?

Yes — plugsky-vision-fast, plugsky-qwen-vl and plugsky-max accept image inputs.

Where is the full docs hub?

/docs — with RAG, embeddings, agents, OAuth, security, and migration guides.