jq-docs-mcp
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@jq-docs-mcp get_price 有哪些参数?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
问题: AI 写聚宽策略时总是瞎编函数签名、搞错参数名、返回值类型全靠猜。
方案: 把聚宽官方文档塞进 AI 的工具箱,查了再写,一次写对。
jq-docs-mcp 是一个 MCP 服务器,内置 221 个 API 函数 和 2,479 个数据表字段 的完整文档。一行命令安装,AI 直接查表写代码。
30 秒安装
Claude Code:
claude mcp add jq-docs -- uvx --from git+https://github.com/jiaweizhang1995/jq-docs-mcp jq-docs-mcp没了。重启 Claude,开始用。
编辑 ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"jq-docs": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/jiaweizhang1995/jq-docs-mcp",
"jq-docs-mcp"
]
}
}
}Related MCP server: Quant Research Platform MCP Server
效果演示
装好之后,直接对 Claude 说人话:
👤 get_price 怎么用?有哪些参数?
🤖 [调用 lookup_function] → 返回完整签名、8个参数说明、返回值类型、示例代码
👤 搜索融资融券相关的 API
🤖 [调用 search_docs] → 找到 get_mtss、get_margincash_stocks 等 4 个 API
👤 balance_sheet 表有哪些字段?
🤖 [调用 lookup_table_columns] → 返回 123 个字段的名称、类型、中文含义
👤 帮我写一个获取茅台近一年日线数据的策略
🤖 [先查 get_price 文档,确认参数] → 写出正确代码,不靠猜没有这个工具: AI 靠记忆编代码 → 参数名拼错 → 运行报错 → 来回改
有了这个工具: AI 先查文档再写 → 一次正确 → 直接跑
6 个查询工具
工具 | 用途 | 示例 |
| 精确查询函数文档 | "get_price 的参数是什么" |
| 中英文关键词搜索 | "搜索融资融券" |
| 按分类浏览 | "获取股票数据分类下有哪些函数" |
| 分类内搜索 | "在财务数据里搜索 balance" |
| 列出全部 221 个函数 | "都有哪些可用的 API" |
| 查询数据表字段 | "income_statement 有哪些列" |
数据覆盖
从聚宽官方 9 个帮助页面完整抓取:
📊 221 个 API 函数 — 签名、参数、返回值、示例代码
📋 2,479 个表字段 — 字段名、类型、中文含义
🔍 支持中英文搜索 — "融资融券" 和 "get_mtss" 都能找到
❌ 找不到时给建议 — 输错函数名会推荐相似函数高级用法
自定义数据库
默认使用内置数据库。如果你自己抓取了更新的数据:
JQ_DB_PATH=/path/to/your/jq_knowledge.db jq-docs-mcp重新抓取
git clone https://github.com/jiaweizhang1995/jq-docs-mcp.git
cd jq-docs-mcp
uv sync
uv run python run_scrape.pyAPI 文档页面公开可访问,无需登录。策略页面需要手机号登录。
技术栈
运行时依赖: 仅 FastMCP — 轻量、零配置
数据库: SQLite 只读模式,757KB,随包分发
传输协议: stdio(本地运行,无需网络)
License
MIT
Available Tools
6 toolslist_by_sectionB
List all API functions in a specific documentation section. 列出某个文档分类下的所有API函数。
Args: section: Section name (e.g. '获取股票数据', '获取融资融券标的列表')
Inputs: section (required). Returns: result
list_functionsA
List all available JoinQuant API function names, grouped by section. 列出所有可用的聚宽API函数名,按文档分类分组。
Use this to discover what functions are available before looking up specific ones.
Inputs: no named parameters. Returns: result
lookup_functionA
Look up complete documentation for a JoinQuant API function by exact name. 查询聚宽API函数的完整文档,包括参数列表、返回值和示例代码。
Args: function_name: Exact function name (e.g. 'get_price', 'get_fundamentals')
Inputs: function_name (required). Returns: result
lookup_table_columnsA
Look up all column definitions for a JoinQuant data table. 查询聚宽数据表的所有字段定义(字段名、类型、含义)。
Critical for get_fundamentals queries where exact column names are needed.
Args: table_name: Table name (e.g. 'balance_sheet', 'income_statement')
Inputs: table_name (required). Returns: result
search_docsA
Search JoinQuant API docs by keyword (English or Chinese). 搜索API文档,支持英文和中文关键词搜索(函数名、中文名、描述、调用签名)。
Args: keyword: Search term in English or Chinese
Inputs: keyword (required). Returns: result
search_in_sectionA
Search API docs within a specific section only. 在指定文档分类中搜索API函数。
Args: keyword: Search term in English or Chinese section: Section name to search within (e.g. '获取股票数据')
Inputs: keyword (required), section (required). Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 6 tools
Each tool has a clear primary purpose, but search_docs and search_in_section overlap as both perform searches; the latter is just a filtered version of the former. All other tools are distinctly about lookup, listing, or column definitions.
All tool names follow a consistent snake_case verb_noun pattern (lookup_function, search_docs, list_by_section, etc.), making the naming predictable and easy to infer.
Six tools is well within the ideal range for a documentation lookup server, covering search, listing, and detail retrieval without unnecessary bloat.
The tool surface covers the full documentation workflow: discover functions via listing/searching, get detailed docs, and retrieve table column definitions. No obvious gaps for the stated purpose.
Related MCP Connectors
Research-only MCP server: your AI as a quant research desk. 90 tools, no trades, no brokers.
Trading MCP server for AI agents, with live market data, account reads and controlled execution.
MCP server for stocksense-ai documentation, generated by doc2mcp.
China A-share market data for research, backtesting and AI agents via MCP.
Related MCP Servers
- AlicenseAqualityDmaintenanceAn MCP server that gives AI assistants direct, structured access to China A-share market data.3Apache 2.0
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to directly access quant research knowledge, including factor libraries, strategy backtesting, and research reports, through the MCP protocol.105MIT
- AlicenseNot gradedqualityDmaintenanceMCP server providing 100% offline access to QuantConnect/Lean documentation with semantic search, code validation, repair, and scaffolding tools for AI coding assistants.2MIT
- FlicenseNot gradedqualityCmaintenanceProvides real-time stock data, technical indicators, and market analysis via MCP protocol to enable AI-driven quantitative trading.1-