響應 API

響應 API 提供了一個有状態的高層接口,用於與 Link Astra 模型交互。它會管理對話状態、支援工具調用,並能在單次請求中执行多步骤的智能體工作流。

POST https://api.linkastra.ai/v1/responses

注意: 響應 API 僅對部分模型開放。兼容性請查閲模型概览

請求體

參數類型必填說明
modelstring模型 ID
inputstring/array用户輸入或對話歷史
instructionsstring系統級指令
toolsarray可用工具(網页搜索、文件搜索、程式碼解釋器、函數)
tool_choicestring/object工具選擇模式
temperaturefloat採样温度(0–2)
max_output_tokensinteger最大輸出 token 數
previous_response_idstring從上一次響應继續
streamboolean通過 SSE 流式返回結果

請求示例

curl https://api.linkastra.ai/v1/responses \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4-pro",
    "input": "What are the key differences between REST and GraphQL?"
  }'

響應

{
  "id": "resp_abc123",
  "object": "response",
  "created_at": 1719000000,
  "model": "deepseek/deepseek-v4-pro",
  "status": "completed",
  "output": [
    {
      "type": "message",
      "id": "msg_abc123",
      "role": "assistant",
      "content": [
        {
          "type": "output_text",
          "text": "REST and GraphQL are two approaches to building APIs..."
        }
      ]
    }
  ],
  "usage": {
    "input_tokens": 15,
    "output_tokens": 200,
    "total_tokens": 215
  }
}
```bash

## 使用 previous_response_id 进行多輪對話

傳入 `previous_response_id` 即可延續對話,無需重新發送完整歷史:

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.linkastra.ai/v1",
    api_key="YOUR_API_KEY"
)

response_1 = client.responses.create(
    model="deepseek/deepseek-v4-pro",
    input="What is machine learning?"
)

response_2 = client.responses.create(
    model="deepseek/deepseek-v4-pro",
    input="How is it different from deep learning?",
    previous_response_id=response_1.id
)

使用工具

響應 API 支援內置工具和自定義函數調用:

response = client.responses.create(
    model="deepseek/deepseek-v4-pro",
    input="What's the weather in Beijing?",
    tools=[{
        "type": "function",
        "name": "get_weather",
        "description": "Get current weather for a location",
        "parameters": {
            "type": "object",
            "properties": {
                "location": {"type": "string"}
            },
            "required": ["location"]
        }
    }]
)
```bash

## 相關文件

- [對話補全](../api-reference/chat-completions) — 更底層的對話 API
- [函數調用](../api-reference/function-calling) — 详細的工具使用指南