Query Table

Query Table is an MCP server that scrapes and extracts structured financial data tables from major Chinese market research platforms using Playwright. It connects AI assistants directly to sources including Tonghuashun iWenCai, TongDaXin WenXiaoDa, and EastMoney stock screening interfaces, alongside conversational AI search platforms like Tencent Yuanbao, Baidu AI Search, and Nami Search. Quantitative researchers, financial analysts, and retail traders use this server to automate equity research, ETF rankings, and industry sector comparisons across different providers with built-in fallback options. By automating browser interactions and intercepting network responses, Query Table avoids complex manual web scraping and API authentication barriers. It outputs tabular data formatted as Markdown, CSV, or pandas DataFrames directly into LLM agent workflows, enabling models to analyze historical stock performance, query market sector caps, and compare fund yields without leaving their chat workspace.

Category: Browser & Web Automation

Tags: data-extraction, financial-data, playwright, web scraping

Visit Query Table

How to install and configure Query Table

1. Install the package via pip: pip install -i https://pypi.org/simple --upgrade mcp_query_table. 2. Ensure Google Chrome is installed and running with remote debugging enabled, for example on http://127.0.0.1:9222. 3. Open your MCP client configuration file (such as Cline or Claude Desktop config). 4. Register the server under mcpServers using the STDIO configuration format: json { "mcpServers": { "mcp_query_table": { "timeout": 300, "command": "python", "args": [ "-m", "mcp_query_table", "--format", "markdown", "--endpoint", "http://127.0.0.1:9222", "--executable_path", "C:\\Program Files\\Google\\Chrome\\Application\\chrome.exe" ] } } } 5. Save the configuration and restart your MCP client.

What you can do with Query Table

• Scrape top-performing ETF lists from Tonghuashun iWenCai by querying natural language trading performance metrics. - Retrieve year-to-date return rankings for mutual funds and equities across TongDaXin query portals. - Extract industry sector valuation tables sorted by circulating market capitalization from Chinese market data sources. - Gather daily top-gaining concept boards and their underlying stock constituents directly from EastMoney screening tools. - Switch data collection fallbacks seamlessly across financial portals when primary web pages change layout or experience downtime.

Key facts

• https://github.com/wukan1986/mcp_query_table
• Browser & Web Automation, Data & Analytics, Finance, Crypto & Payments
• data-extraction, financial-data, playwright, web scraping

Part of MCP Servers

How do I install Query Table?

You can install Query Table using pip by running pip install -i https://pypi.org/simple --upgrade mcp_query_table. You also need a local installation of Google Chrome capable of starting with remote debugging enabled, allowing Playwright to control the browser session.

What can Query Table do?

Query Table allows AI agents to query and retrieve structured financial tables from financial portals like Tonghuashun, TongDaXin, and EastMoney. It captures network responses to extract ETF rankings, stock screeners, and sector returns, returning data formatted as Markdown or CSV.

Which MCP clients work with Query Table?

Query Table works with any standard Model Context Protocol client that supports STDIO, SSE, or Streamable HTTP connections, including Cline, Claude Desktop, and OtterMind. It also supports debugging using the official MCP Inspector tool.

Why does Query Table require a higher timeout setting?

Browser automation involves launching pages, executing user interactions, waiting for scripts to execute, and navigating through table pagination. These operations frequently take longer than simple API requests, requiring timeout values of 300 seconds in MCP client settings.

Does Query Table require user account login?

Query Table does not handle automated login workflows. If a target financial website requires an authenticated user session for pagination or detailed queries, you must log into your account manually in the designated browser profile before running queries.