FinData is an open-source Model Context Protocol (MCP) server designed to provide large language models (LLMs) with professional access to financial data. It supports multiple data providers, including Tushare, Wind, DataYes, and others, enabling AI applications to quickly retrieve financial data.
Main capabilities:
- Multi-data provider support: Specify the data source by setting the PROVIDER environment variable. Currently, Tushare, Wind, DataYes, and others are supported.
- Flexible transport methods: Fully supports both Stdio and SSE transport methods, adapting to different deployment environments. The Stdio method requires editing the MCP client configuration file, while the SSE method can specify the binding host and port (default localhost:8000).
- Rich financial data interfaces:
- Stock daily quotes (daily): Retrieve unadjusted daily market data.
- Fundamental data: Includes basic stock information (stock_basic), listed company basic information (stock_company), stock fundamental data within a specified time range (bak_basic), income statement (income), balance sheet (balancesheet), cash flow statement (cashflow).
- Macroeconomic data: Includes Loan Prime Rate (shibor_lpr), Gross Domestic Product (cn_gdp), Consumer Price Index (cn_cpi), Producer Price Index (cn_ppi), money supply (cn_m), social financing (sf_month), Purchasing Managers' Index (cn_pmi).
Usage:
- Preparation: Install mcp[cli]>=1.6.0, and install optional packages according to the data provider, such as tushare>=1.4.21.
- Configuration: For Stdio transport, edit the MCP client configuration file to add finData, and set the environment variables DATA_API_TOKEN and PROVIDER.
- Start the MCP server: Optionally use the parameters --sse-host and --sse-port to specify the host and port of the SSE server (default localhost:8000).
- Connect: After the MCP server is running, update the MCP client configuration with the corresponding settings to connect. Note that the configuration file variable names may differ slightly among different MCP clients; please refer to each client's documentation for configuration.
Applicable scenarios:
- AI agents need real-time or historical financial data for analysis, report generation, or decision support.
- Develop LLM-based financial applications, such as investment research assistants, financial analysis tools, etc.
- Complex scenarios that require integrating multiple financial data sources.
Limitations:
- Before use, you must have the API token (DATA_API_TOKEN) of the corresponding data provider.
- Specific data fields and return formats should refer to the corresponding provider's documentation; this project may not provide all details.
- Additional information regarding performance, pricing, free quotas, etc., is not currently provided.
This project is open-sourced by DataCanvas. For more information, please visit the GitHub repository:https://github.com/zlinzzzz/finData-mcp-server
