Chat Completions

API

Use the OpenAI-compatible Chat Completions API for text, vision, tools, and streaming.

Overview

Chat Completions is the primary endpoint for conversational model calls. It accepts a list of `messages` and returns an assistant response. When `stream` is `true`, output is delivered as Server-Sent Events.

Request

`POST https://uouo.cloud/v1/chat/completions`

Headers

Header	Required	Description
Authorization	Yes	`Bearer $UOUODUO_API_KEY`
Content-Type	Yes	`application/json`

Body parameters

Parameter	Type	Required	Default	Description
model	string	Yes	-	Model ID, for example `gpt-4o-mini`.
messages	array	Yes	-	Conversation messages.
temperature	number	No	1	Sampling temperature from 0 to 2.
top_p	number	No	1	Nucleus sampling value from 0 to 1.
stream	boolean	No	false	Return SSE chunks when true.
max_tokens	integer	No	-	Legacy maximum output token field.
max_completion_tokens	integer	No	-	Maximum generated tokens.
tools	array	No	-	Function/tool definitions.
tool_choice	string or object	No	`auto`	Controls tool selection.
response_format	object	No	-	JSON object or schema output constraints.
user	string	No	-	End-user identifier for audit and abuse prevention.

Example

curl https://uouo.cloud/v1/chat/completions \
  -H "Authorization: Bearer $UOUODUO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [
      { "role": "system", "content": "You are a concise production assistant." },
      { "role": "user", "content": "Summarize the release risk in three bullets." }
    ]
  }'

Streaming

Set `"stream": true` to receive `text/event-stream` chunks:

data: {"id":"chatcmpl_abc","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"content":"First"},"finish_reason":null}]}

data: {"id":"chatcmpl_abc","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"content":" point"},"finish_reason":null}]}

data: [DONE]

If `stream_options.include_usage` is enabled, the final chunk may include usage. Network interruptions can prevent the final usage chunk from arriving, so use console logs as the billing source of truth.

Response fields

Field	Type	Description
id	string	Completion ID.
object	string	Usually `chat.completion`.
created	integer	Unix timestamp.
model	string	Model used by the route.
choices	array	Candidate responses.
usage	object	Prompt, completion, and total token counts.

Errors

HTTP	Code	Meaning	Suggested action
400	invalid_request_error	Invalid body or parameter range.	Validate the request schema.
401	unauthorized	Missing or invalid API key.	Create or rotate a key in `/app/keys`.
403	forbidden	Key or account cannot access the model.	Check model access and workspace policy.
429	rate_limit_exceeded	Rate limit, quota, or balance issue.	Review usage and retry with backoff.
500	internal_error	Gateway or upstream failure.	Record request ID and check status.

Notes

Vision input requires a model that supports image content parts.
Tool call arguments should be validated by your server before execution.
Provider support for `seed`, `reasoning_effort`, and structured output can vary.
Control cost with `max_completion_tokens` and monitor `/app/usage`.