推荐 endpoint
最小请求
{
"model": "gpt-4.1",
"input": "从这段文本中提取订单号和金额:订单 A-1024,金额 99.5 美元。",
"text": {
"format": {
"type": "json_schema",
"name": "order",
"strict": true,
"schema": {
"type": "object",
"properties": {
"order_id": { "type": "string" },
"amount": { "type": "number" }
},
"required": ["order_id", "amount"],
"additionalProperties": false
}
}
}
}
cURL 示例
curl https://mass.apigo.ai/v1/responses \
-H "Authorization: Bearer $TIDEMIND_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-4.1",
"input": "从这段文本中提取订单号和金额:订单 A-1024,金额 99.5 美元。",
"text": {
"format": {
"type": "json_schema",
"name": "order",
"strict": true,
"schema": {
"type": "object",
"properties": {
"order_id": { "type": "string" },
"amount": { "type": "number" }
},
"required": ["order_id", "amount"],
"additionalProperties": false
}
}
}
}'
Python 示例
import json
from openai import OpenAI
client = OpenAI(
base_url="https://mass.apigo.ai/v1",
api_key="<TIDEMIND_API_KEY>",
)
response = client.responses.create(
model="gpt-4.1",
input="从这段文本中提取订单号和金额:订单 A-1024,金额 99.5 美元。",
text={
"format": {
"type": "json_schema",
"name": "order",
"strict": True,
"schema": {
"type": "object",
"properties": {
"order_id": {"type": "string"},
"amount": {"type": "number"},
},
"required": ["order_id", "amount"],
"additionalProperties": False,
},
}
},
)
print(json.loads(response.output_text))
Node.js 示例
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://mass.apigo.ai/v1",
apiKey: process.env.TIDEMIND_API_KEY,
});
const response = await client.responses.create({
model: "gpt-4.1",
input: "从这段文本中提取订单号和金额:订单 A-1024,金额 99.5 美元。",
text: {
format: {
type: "json_schema",
name: "order",
strict: true,
schema: {
type: "object",
properties: {
order_id: { type: "string" },
amount: { type: "number" }
},
required: ["order_id", "amount"],
additionalProperties: false
}
}
}
});
console.log(JSON.parse(response.output_text));
最佳实践
- JSON Schema 尽量小而硬,不要一上来就做深层嵌套
- 用
strict: true和additionalProperties: false降低漂移 - 服务端仍要做二次校验,不要直接信任模型输出
