跳转到主要内容
POST
/
v1
/
chat
/
completions
curl --request POST \ --url https://api.shuyou.ai/v1/chat/completions \ --header 'Authorization: Bearer <token>' \ --header 'Content-Type: application/json' \ --data ' { "model": "openai/gpt-5.5-pro", "messages": [ { "role": "developer", "content": "You are a helpful assistant." }, { "role": "user", "content": "Hello!" } ] } '
{
  "id": "chatcmpl-example",
  "object": "chat.completion",
  "created": 1741569952,
  "model": "openai/gpt-5.5-pro",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello! How can I assist you today?",
        "refusal": null,
        "annotations": []
      },
      "logprobs": null,
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 19,
    "completion_tokens": 10,
    "total_tokens": 29,
    "prompt_tokens_details": {
      "cached_tokens": 0,
      "audio_tokens": 0
    },
    "completion_tokens_details": {
      "reasoning_tokens": 0,
      "audio_tokens": 0,
      "accepted_prediction_tokens": 0,
      "rejected_prediction_tokens": 0
    }
  }
}

授权

​
Authorization
string
header
必填

Authorization: Bearer YOUR_API_KEY

请求体

application/json
​
messages
object[]
必填

A list of messages comprising the conversation so far. Supports text, images, audio, and file content parts depending on the model.

Minimum array length: 1
​
model
string
必填

Model ID (provider/model_name). Use the slug from List Models.

示例:

"openai/gpt-5.5-pro"

​
max_completion_tokens
integer | null

Upper bound for generated tokens, including visible output and reasoning tokens.

​
temperature
number
默认值:1

Sampling temperature. Higher values are more random; lower values are more deterministic.

必填范围: 0 <= x <= 2
​
top_p
number
默认值:1

Nucleus sampling: only tokens within the top_p probability mass are considered.

必填范围: 0 <= x <= 1
​
n
integer
默认值:1

Number of chat completion choices per input message. ShuYou currently supports n=1 only.

必填范围: 1 <= x <= 128
​
frequency_penalty
number
默认值:0

Penalize tokens based on existing frequency in the text so far.

必填范围: -2 <= x <= 2
​
presence_penalty
number
默认值:0

Penalize tokens based on whether they appear in the text so far.

必填范围: -2 <= x <= 2
​
stop

Up to 4 sequences where the API stops generating. Not supported on some reasoning models.

​
logit_bias
object

Maps token IDs to bias values (-100 to 100) applied before sampling.

​
logprobs
boolean | null
默认值:false

Return log probabilities of output tokens.

​
top_logprobs
integer

Number of most likely tokens per position. Requires logprobs: true.

必填范围: 0 <= x <= 20
​
tools
object[]

Tools the model may call (function or custom tools).

​
tool_choice

Controls which tool(s) the model calls. Default: auto when tools are present, else none.

可用选项:
none,
auto,
required
​
parallel_tool_calls
boolean
默认值:true

Whether to allow parallel tool calls in a single response.

​
reasoning_effort
enum<string>

Reasoning effort for reasoning models. Supported values vary by model.

可用选项:
none,
minimal,
low,
medium,
high,
xhigh
​
verbosity
enum<string>

Constrains response verbosity.

可用选项:
low,
medium,
high
​
web_search_options
object

Configuration for the web search tool.

​
metadata
object

Up to 16 key-value pairs (keys max 64 chars) for structured metadata.

​
response_format
object

Output format: text, json_object, or structured json_schema.

​
stream
boolean | null
默认值:false

Stream the response via server-sent events.

​
stream_options
object

Options when stream is true.

​
reasoning
object

ShuYou: configure reasoning trace behavior.

​
provider
object

ShuYou: multi-provider routing configuration.

​
model_routing_config
object

ShuYou: model selection within a provider.

响应

Chat completion object, or SSE stream when stream is true

​
id
string
​
object
enum<string>
可用选项:
chat.completion
​
created
integer

Unix timestamp (seconds).

​
model
string
​
choices
object[]
​
usage
object
​
system_fingerprint
string