home / mcp / financemcp mcp server

FinanceMCP MCP Server

Provides real-time financial data, indicators, and market analyses via MCP protocols across global markets.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "guangxiangdebizi-financemcp": {
      "url": "https://finvestai.top/mcp",
      "headers": {
        "TUSHARE_TOKEN": "YOUR_TUSHARE_TOKEN",
        "X-Tushare-Token": "YOUR_TUSHARE_TOKEN_HERE"
      }
    }
  }
}

You can run FinanceMCP, a professional financial data MCP server, to access real-time market data, indicators, and financial analysis through MCP-compatible clients. It supports public cloud HTTP access and local SSE/stdio deployments, enabling you to plug powerful financial data into AI assistants or custom tooling with minimal setup.

How to use

Connect to FinanceMCP using an MCP client in either HTTP (remote) mode or local stdio mode. In HTTP mode, your client will reach the server at the defined URL and pass tokens via headers or tokens as configured. In stdio mode, your local client launches the MCP server as a standalone process and communicates through standard I/O, enabling fast, offline-capable operation. Once connected, you can request stock data, indicators, macro data, market news, and company analyses across multiple markets.

How to install

Prerequisites you need before installing FinanceMCP include Node.js version 18 or newer and a Tushare API token if you plan to access paid data sources.

# Step 1: Install the runtime client (stdio mode) or run via npx
# If you plan to use the stdio mode with a local installation:
npm install -g finance-mcp

# Step 2: Prepare your Tushare token (if required by your setup)
# Obtain your token from tushare.pro and keep it handy
export TUSHARE_TOKEN=YOUR_TUSHARE_TOKEN

# Step 3: Start the stdio server (example using CLI after global install)
finance-mcp

Configuration and connection options

FinanceMCP can be connected via two explicit methods shown in the setup examples. Use the HTTP method for remote cloud deployment and the stdio method for local, zero-config operation. Both methods support the same MCP endpoints and tools.

{
  "mcpServers": {
    "finance-mcp": {
      "type": "stdio",
      "command": "finance-mcp",
      "args": [],
      "env": {
        "TUSHARE_TOKEN": "YOUR_TUSHARE_TOKEN"
      }
    },
    "finance-data-server": {
      "type": "http",
      "url": "https://finvestai.top/mcp",
      "args": [],
      "headers": {
        "X-Tushare-Token": "YOUR_TUSHARE_TOKEN"
      }
    }
  }
}

Inline steps to use the two deployment modes

For local, zero-setup usage, run the stdio-based server and configure your MCP client to connect via the provided runtime command. For remote access, point your client at the HTTP endpoint and include the token in the headers as shown.

Available tools

current_timestamp

Return the current timestamp in UTC+8 with multiple output formats.

finance_news

Search financial news with keyword queries.

stock_data

Retrieve stock or crypto data across markets and compute technical indicators.

index_data

Provide index data for major markets.

csi_index_constituents

Summarize CSI index constituents, weights, and valuation metrics.

macro_econ

Deliver macroeconomic indicators such as GDP, CPI, PPI, PMI.

company_performance

Analyze A-share company finances, management, and fundamentals.

company_performance_hk

Analyze Hong Kong-listed company finances.

company_performance_us

Analyze US-listed company finances.

fund_data

Fund data including net value, holdings, and distributions.

fund_manager_by_name

Query fund managers and their portfolios.

convertible_bond

Provide information on convertible bonds.

block_trade

Access large block trade data.

money_flow

Analyze capital flow across stocks, markets, and sectors.

margin_trade

Access financing and margin-related data.

dragon_tiger_inst

Fetch龙虎榜 institutional details for specific dates.

hot_news_7x24

7x24 hot news with deduplication and cross-source similarity handling.