Xueqiu MCP
# Xueqiu MCP
基于雪球API的MCP服务，让您通过Claude或其他AI助手轻松获取股票数据。
## 项目简介
本项目基于[pysnowball](https://github.com/uname-yang/pysnowball)封装了雪球API，并通过MCP协议提供服务，使您能够在Claude等AI助手中直接查询股票数据。
## 安装方法
本项目使用`uv`进行依赖管理。请按照以下步骤进行安装：
```bash
# 克隆仓库
git clone https://github.com/liqiongyu/xueqiu_mcp.git
cd xueqiu_mcp
# 使用uv安装依赖
uv venv && uv pip install -e .
```
## 配置
### 配置雪球Token
1. 在项目根目录创建`.env`文件
2. 添加以下内容：
```
XUEQIU_TOKEN=您的雪球token
```
* 快捷方式：
```bash
echo 'XUEQIU_TOKEN="xq_a_token=xxxxx;u=xxxx"' > .env
```
关于如何获取雪球token，请参考[pysnowball文档](https://github.com/uname-yang/pysnowball/blob/master/how_to_get_token.md)。
## 运行服务
使用以下命令启动MCP服务：
```bash
uv --directory /path/to/xueqiu_mcp run main.py
```
或者，如果您已经配置了Claude Desktop：
```json
"xueqiu-mcp": {
"args": [
"--directory",
"/path/to/xueqiu_mcp",
"run",
"main.py"
],
"command": "uv"
}
```
## 功能特性
- 获取股票实时行情
- 查询指数收益
- 获取深港通/沪港通北向数据
- 基金相关数据查询
- 关键词搜索股票代码
## 展示图
## 致谢
- [pysnowball](https://github.com/uname-yang/pysnowball) - 雪球股票数据接口的Python版本
- [fastmcp](https://github.com/fastmcp) - MCP服务框架
## 许可证
[MIT License](./LICENSE)TDQS
Scored across 45 tools
Most tools have distinct purposes targeting specific financial data types (e.g., balance sheet, cash flow, fund info), but there is some overlap in tools like quotec and quote_detail for stock quotes, and multiple fund tools (fund_info, fund_detail) could be confusing. The descriptions help differentiate, but the sheer number increases potential for misselection.
Naming is mostly consistent with snake_case and descriptive nouns (e.g., capital_flow, income, fund_manager), with clear patterns for related tools (e.g., index_perf_30, index_perf_90). Minor deviations exist, such as pankou (non-English) and quotec (abbreviated), but overall the conventions are readable and predictable.
With 45 tools, the count is excessive for a financial data server, making it heavy and potentially overwhelming. While the domain is broad (stocks, funds, indices, etc.), many tools could be consolidated (e.g., multiple fund tools, index performance periods), indicating poor scoping and unnecessary fragmentation.
The tool set covers a wide range of financial data, including stocks, funds, indices, and market metrics, with good CRUD-like coverage for key entities (e.g., financial statements, fund details). Minor gaps exist, such as limited update/delete operations typical for read-only data, but agents can likely access most needed information without major dead ends.