API 接入

curl -X POST https://your-domain/v1/search \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer ais-xxxxxxxx' \
  -d '{
    "query": "golang gin tutorial",
    "max_results": 8,
    "include_content": false,
    "topic": "general",
    "freshness": "week",
    "language": "en"
  }'

from openai import OpenAI
client = OpenAI(base_url="https://your-domain/v1", api_key="ais-xxxxxxxx")
resp = client.chat.completions.create(
    model="any",                      # 未配置 LLM 时 model 仅为占位
    messages=[{"role": "user", "content": "gin 框架最新版本有什么特性？"}],
)
print(resp.choices[0].message.content)  # 返回带引用链接的搜索结果

{
  "mcpServers": {
    "seekcapy": {
      "type": "http",
      "url": "https://your-domain/mcp",
      "headers": { "Authorization": "Bearer ais-xxxxxxxx" }
    }
  }
}

curl -X POST https://your-domain/v1/rerank \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer ais-xxxxxxxx' \
  -d '{
    "query": "什么是 RAG 检索增强生成？",
    "documents": [
      "检索增强生成先检索再生成…",
      "今天的午餐特价是牛排沙拉。"
    ],
    "top_n": 2
  }'

import requests

resp = requests.post("https://your-domain/v1/rerank",
    headers={"Authorization": "Bearer ais-xxxxxxxx"},
    json={
        "query": "What is RAG?",
        "documents": [
            "RAG retrieves documents first, then generates answers.",
            "Today lunch special is steak salad.",
            "Vector databases store embeddings for semantic search.",
        ],
        "top_n": 2,
    },
    timeout=30,
).json()

for r in resp["data"]["results"]:
    print(r["index"], round(r["relevance_score"], 4))

{
  "code": 0,
  "data": {
    "results": [
      { "index": 0, "relevance_score": 0.86 },
      { "index": 2, "relevance_score": 0.41 }
    ],
    "model": "seekcapy-rerank-bge-base",
    "took_ms": 12,
    "documents_count": 3,
    "billing_units": 1
  }
}