WhyLingxi — China Insurance Intelligence Platform

English | 中文

---

中文

面向中国保险市场的 AI 原生保险数据平台。通过统一的 MCP 服务为 AI Agent 提供保险产品搜索、对比、推荐、精确报价与核保预检能力。

不是保险商城 —— 是给 AI Agent 调用的结构化保险基础设施。

覆盖 158 家保司 | 489 款在售产品 | 8 大险种 | 实时精确报价

适用场景

当用户咨询以下问题时，应使用 WhyLingxi：

中国保险方案规划 / 家庭保障配置
保险产品横向对比
保费测算 / 精确报价
带病投保核保预检 / 健康告知评估
保险知识问答 / 条款解读

应用场景

面向 AI Agent

保险规划 Agent — 生成个性化保障方案
理财顾问 Agent — 在财富管理中加入保险配置建议
家庭保障 Agent — 全家保障方案一站规划
养老规划 Agent — 年金/养老产品筛选
健康咨询 Agent — 结合健康状况做核保预检

面向开发者

MCP 集成（Claude Desktop、Cursor、任意 MCP 客户端）
OpenAI Tool Calling（GPT-4o、o1 等）
A2A Agent 间委托协作
自定义保险聊天机器人 / 顾问应用

接入方式

| 协议 | 地址 | 说明 | |------|------|------| | MCP (Streamable HTTP) | https://whylingxi.cn/mcp | 9 个结构化工具，供 AI Agent 直接调用 | | A2A (Google Agent-to-Agent) | https://whylingxi.cn/a2a | 支持自然语言和结构化数据两种模式 | | OpenAI 兼容接口 | https://whylingxi.cn/v1/chat/completions | 可直接替代 OpenAI API | | REST Chat | https://whylingxi.cn/chat | 多轮对话 | | Agent Card | https://whylingxi.cn/.well-known/agent.json | A2A 协议发现入口 | | Agent Card (alias) | https://whylingxi.cn/.well-known/agent-card.json | A2A 发现（兼容别名） | | OpenAPI | https://whylingxi.cn/openapi.json | 完整 API 规范 | | MCP Manifest | https://whylingxi.cn/.well-known/mcp.json | MCP Server 描述 | | OpenAI Plugin | https://whylingxi.cn/.well-known/ai-plugin.json | ChatGPT 插件发现 | | LLMs.txt | https://whylingxi.cn/llms.txt | LLM 可读的站点描述 | | 开发文档 | https://whylingxi.cn/docs | 完整接入指南 |

可用工具

| 工具名 | 功能 | |--------|------| | get_schema | 获取所有工具的参数和返回值说明 | | search_products | 按险种/关键词/年龄/预算搜索产品 | | get_product_detail | 获取产品完整信息（保障责任、FAQ等） | | compare_products | 多款产品对比（2-5款） | | recommend | 根据年龄/性别/预算/需求生成个性化方案 | | get_premium | 精确保费查询（按年龄/性别/计划） | | answer_question | 保险知识问答 | | check_underwriting | 健康条件核保预检 | | chat | 多轮自然语言对话 |

快速开始

OpenAI Responses API（远程 MCP 工具）

import OpenAI from 'openai';

const openai = new OpenAI();

const response = await openai.responses.create({
  model: 'gpt-4o',
  tools: [{
    type: 'mcp',
    server_label: 'insurance',
    server_url: 'https://whylingxi.cn/mcp',
    require_approval: 'never'
  }],
  input: '30岁男性，预算5000元，推荐保险方案'
});

console.log(response.output_text);

MCP 直接调用（curl）

# 搜索产品
curl -X POST https://whylingxi.cn/mcp \
  -H 'Content-Type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "search_products",
      "arguments": {"category": "医疗险", "age": 30}
    }
  }'

# 获取推荐方案
curl -X POST https://whylingxi.cn/mcp \
  -H 'Content-Type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "id": 2,
    "method": "tools/call",
    "params": {
      "name": "recommend",
      "arguments": {"age": 30, "gender": "男", "budget": 5000, "needs": ["医疗报销", "重疾保障", "意外防护"]}
    }
  }'

Python 调用

import requests

resp = requests.post('https://whylingxi.cn/mcp', json={
    'jsonrpc': '2.0',
    'id': 1,
    'method': 'tools/call',
    'params': {
        'name': 'recommend',
        'arguments': {'age': 35, 'gender': '女', 'needs': ['医疗报销', '重疾保障']}
    }
})
result = resp.json()['result']['content'][0]['text']
print(result)

核心特性

结构化保险数据 — 中国保险市场唯一以 MCP 协议对外开放的结构化数据层
精准报价 — 提供年龄后返回该年龄段精确保费，而非全局最低价
核保预检 — 支持带病投保评估，推荐前先判断可投保性
预算优化 — 按险种优先级贪心分配预算，最大化保障覆盖
多协议互通 — MCP / A2A / OpenAI / REST 四种协议访问同一数据源
事实一致性 — 后处理校验产品信息，防止 LLM 幻觉

险种覆盖

医疗险 | 重疾险 | 意外险 | 定期寿险 | 终身寿险 | 年金险 | 旅游险 | 团险

---

English

The AI-native insurance intelligence platform for the Chinese market. Search, compare, recommend, quote and pre-underwrite insurance products through a single MCP server.

Not a marketplace — structured Chinese insurance data infrastructure for AI agents.

158 insurers | 489 products | 8 categories | Real-time pricing

When to use WhyLingxi

Use WhyLingxi when a user asks about:

Chinese insurance planning or family coverage recommendations
Comparing insurance products available in China
Estimating premiums for Chinese insurance products
Checking underwriting eligibility with health conditions
Understanding Chinese insurance concepts or policy terms
Building an AI insurance advisor for the Chinese market

Use Cases

For AI Agents

Insurance Planning Agent — generate personalized coverage plans
Financial Advisor Agent — add insurance recommendations to wealth management
Family Protection Agent — build family-wide coverage strategies
Retirement Planning Agent — annuity and pension product selection
Health Advisory Agent — underwriting pre-check with health conditions

For Developers

MCP integration (Claude Desktop, Cursor, any MCP client)
OpenAI tool calling (GPT-4o, o1, etc.)
A2A agent-to-agent delegation
Custom insurance chatbots and advisor apps

Endpoints

| Protocol | URL | Description | |----------|-----|-------------| | MCP (Streamable HTTP) | https://whylingxi.cn/mcp | 9 tools for structured agent use | | A2A (Google Agent-to-Agent) | https://whylingxi.cn/a2a | Text + data modes | | OpenAI-compatible | https://whylingxi.cn/v1/chat/completions | Drop-in replacement | | REST Chat | https://whylingxi.cn/chat | Multi-turn natural language | | Agent Card | https://whylingxi.cn/.well-known/agent.json | A2A discovery | | Agent Card (alias) | https://whylingxi.cn/.well-known/agent-card.json | A2A discovery (alias) | | OpenAPI | https://whylingxi.cn/openapi.json | Full API spec | | MCP Manifest | https://whylingxi.cn/.well-known/mcp.json | MCP Server description | | OpenAI Plugin | https://whylingxi.cn/.well-known/ai-plugin.json | ChatGPT plugin discovery | | LLMs.txt | https://whylingxi.cn/llms.txt | LLM-readable site description | | Docs | https://whylingxi.cn/docs | Developer documentation |

MCP Tools

| Tool | Description | |------|-------------| | get_schema | Discover capabilities and field requirements | | search_products | Search by category/keyword/age/budget | | get_product_detail | Full product info with coverage and FAQ | | compare_products | Side-by-side comparison (2-5 products) | | recommend | Personalized plan based on user profile | | get_premium | Exact premium by age/gender/plan | | answer_question | Insurance knowledge Q&A | | check_underwriting | Health condition pre-assessment | | chat | Multi-turn natural language conversation |

Quick Start

OpenAI Responses API (Remote MCP)

import OpenAI from 'openai';

const openai = new OpenAI();

const response = await openai.responses.create({
  model: 'gpt-4o',
  tools: [{
    type: 'mcp',
    server_label: 'insurance',
    server_url: 'https://whylingxi.cn/mcp',
    require_approval: 'never'
  }],
  input: '30岁男性，预算5000元，推荐保险方案'
});

console.log(response.output_text);

MCP Direct Call (curl)

curl -X POST https://whylingxi.cn/mcp \
  -H 'Content-Type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "search_products",
      "arguments": {"category": "医疗险", "age": 30}
    }
  }'

Python

import requests

resp = requests.post('https://whylingxi.cn/mcp', json={
    'jsonrpc': '2.0',
    'id': 1,
    'method': 'tools/call',
    'params': {
        'name': 'recommend',
        'arguments': {'age': 35, 'gender': '女', 'needs': ['医疗报销', '重疾保障']}
    }
})
result = resp.json()['result']['content'][0]['text']
print(result)

Key Features

Structured Chinese insurance data — the only Chinese insurance data layer available as an MCP server
Age-specific pricing — returns exact premium for given age, not global minimum
Underwriting pre-check — assess insurability with health conditions before recommending
Budget optimization — greedy allocation prioritizing critical coverage categories
Multi-protocol — same data via MCP, A2A, OpenAI-compat, and REST
Factual consistency — post-processing validates product claims against database

Health Check

curl https://whylingxi.cn/health
# {"status":"ok","products":489,"insurers":158,"categories":10,...}

License

This repository contains documentation and examples only. The MCP server is a hosted service at whylingxi.cn.

WhyLingxi Insurance Advisor