Tushare MCP

Tushare MCP is an MCP server that connects AI assistants like Claude Desktop and Cursor to financial market data from Tushare Pro via the Tinyshare SDK. Built on Python and FastAPI, this tool allows financial analysts, quant developers, and traders to query Chinese A-share and Hong Kong stock markets directly through conversational interfaces. The server provides access to over thirty specialized endpoints, encompassing basic corporate details, trade calendars, daily and weekly price quotes, PE and PB valuation multiples, and dividend distributions. Users can also analyze comprehensive company financials, including income statements, balance sheets, cash flow reports, performance forecasts, and core business segment breakdowns. In addition to equities, Tushare MCP supports mutual funds and exchange-traded funds, providing ETF baseline specs, manager biographies, fund net asset values, historical portfolio holdings, and technical factor calculations like MACD and RSI. The server incorporates a local request batching mechanism that intercepts comma-separated symbol lists and aggregates responses, accommodating standard model prompting habits. By operating locally or over remote endpoints, it provides structured financial context for market research, investment screening, and automated trading workflow design.

Category: Data & Analytics

Tags: financial-data, market data, stocks, tushare

Visit Tushare MCP

How to install and configure Tushare MCP

1. Clone the repository: bash git clone https://github.com/buuzzy/tushare_MCP tushare_mcp cd tushare_mcp 2. Create and activate a Python 3.10+ virtual environment: bash python3 -m venv venv source venv/bin/activate 3. Install the dependencies: bash pip install -r requirements.txt 4. Configure your token in a .env file: bash echo "TUSHARE_TOKEN=your_token" >> .env 5. To use with Cursor, start the server using python server.py. In Cursor Settings > Features > MCP, add a server named tushare with Type set to Streamable HTTP and URL http://localhost:8000/mcp. 6. To use with Claude Desktop, register the server in claude_desktop_config.json with the command pointing to your virtual environment Python path and args set to ["/path/to/server.py", "--stdio"].

What you can do with Tushare MCP

• Querying historical daily, weekly, or monthly price quotes and trading volumes across multiple A-share stocks simultaneously using comma-separated ticker symbols. - Retrieving corporate financial statements, including balance sheets, income statements, and cash flow reports, to evaluate fundamental company health in Claude Desktop. - Screening public mutual funds and ETFs by checking portfolio asset holdings, historical net asset values, fund manager resumes, and technical indicators. - Inspecting daily market valuation metrics such as price-to-earnings, price-to-book ratios, turnover rates, and market capitalization directly within Cursor. - Tracking Northbound and Southbound capital movements through Shanghai-Shenzhen-Hong Kong Stock Connect daily top-ten active trading listings.

Key facts

• https://github.com/buuzzy/tushare_MCP
• Data & Analytics, Finance, Crypto & Payments
• financial-data, market data, stocks, tushare

Part of MCP Servers

What is Tushare MCP?

Tushare MCP is a Model Context Protocol server that enables AI clients like Claude Desktop and Cursor to retrieve Chinese financial market data. It interfaces with Tushare Pro via the Tinyshare SDK, allowing users to query A-share equities, ETF information, fund portfolios, and corporate financial disclosures through natural language prompts.

Which MCP clients work with Tushare MCP?

Tushare MCP works with Claude Desktop and Cursor IDE. Claude Desktop connects locally via the standard Stdio communication mode, while Cursor connects over local or remote Streamable HTTP endpoints.

How does Tushare MCP handle multi-stock queries?

Because standard Tushare fund endpoints do not natively support comma-separated multi-code queries, Tushare MCP includes an internal request splitter and merger. It automatically intercepts batch requests, issues individual API calls sequentially in the background, and concatenates the results into a single dataset for the model.

Is Tushare MCP open source?

Yes, Tushare MCP is open-source software distributed under the MIT License. The code repository, documentation, and underlying tool definitions are publicly accessible on GitHub.

Do I need an API token to use Tushare MCP?

Yes, running your own local instance requires a Tushare Pro API token configured in a local environment file or registered using the built-in token setup tool. The project also provides public trial SSE endpoints for testing without an immediate token setup.