Embeddings API：文本向量

POST /v1/embeddings，与 OpenAI 格式一致，常用于 RAG、语义检索与聚类。

请求

bash

curl "$BASE_URL/v1/embeddings" \
  -H "Authorization: Bearer $NEW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"<向量模型ID>","input":["第一段","第二段"]}'

python

from openai import OpenAI

client = OpenAI(base_url="https://api.nowhub.org/v1", api_key="sk-...")
resp = client.embeddings.create(
    model="<向量模型ID>",
    input=["第一段文本", "第二段文本"],
)
vectors = [d.embedding for d in resp.data]

响应

json

{
  "object": "list",
  "data": [{ "object": "embedding", "index": 0, "embedding": [0.0023, -0.0091] }],
  "usage": { "prompt_tokens": 8, "total_tokens": 8 }
}

注意

按 index 对齐：批量返回顺序不保证与输入一致
维度：换模型必须重建向量索引
可用性：不是所有分组都开了向量模型，先在模型广场确认