POST /v1/chat/completions

Create a model response for a chat conversation. OpenAI-compatible; works with any catalog model.

Request body

Field	Type	Required	Description
model	string	required	Catalog model id, e.g. claude-opus-4-7.
messages	array	required	Conversation turns with role and content.
stream	boolean		Stream tokens over SSE. Default false.
tools	array		Tool definitions for function calling.
max_tokens	integer		Cap on generated tokens; defaults per model.
temperature	number		Sampling temperature; provider defaults apply.

Example

Request

curl https://gateway.oya.moe/v1/chat/completions \
  -H "Authorization: Bearer $ARKAFLOW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.1",
    "stream": true,
    "messages": [
      {"role": "user", "content": "Summarize this…"}
    ]
  }'

Response (SSE)

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

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

data: [DONE]

Errors

Code	Meaning
401 UNAUTHORIZED	Missing or invalid API key.
402 INSUFFICIENT_BALANCE	Wallet empty — top up to resume.
404 NOT_FOUND	Unknown model id — check the catalog.
429 LIMIT_EXCEEDED	Rate limit hit; honor Retry-After.

See Errors for the full vocabulary and retry guidance.