MCP 服务器

Finance Tools MCP

An MCP server that provides comprehensive financial insights and analysis by leveraging real-time market data, news, and advanced analytics for stocks, options, financial statements, and economic indicators.

Tools

get_ticker_data

Get comprehensive report for ticker: overview, news, metrics, performance, dates, analyst recommendations, and upgrades/downgrades.

get_options

Get options with highest open interest. Dates: YYYY-MM-DD. Type: C=calls, P=puts.

get_price_history

Get historical price data digest for specified period. Usually get at least 3 months, 6 months or more.

get_financial_statements

Get financial statements. Types: income, balance, cash. Frequency: quarterly, annual.

get_institutional_holders

Get major institutional and mutual fund holders.

get_earnings_history

Get earnings history with estimates and surprises.

get_insider_trades

Get recent insider trading activity.

get_ticker_news_tool

For getting yahoo financial news of a ticker. Useful for getting latest news, especially for doing deep research.

get_current_fng_tool

Get the current CNN Fear & Greed Index. Returns: str: The current Fear & Greed Index with score and rating.

get_historical_fng_tool

Get historical CNN Fear & Greed Index data for a specified number of days. Parameters: days (int): Number of days of historical data to retrieve (limited by the API). Returns: str: Historical Fear & Greed Index values for the specified period.

analyze_fng_trend

Analyze trends in CNN Fear & Greed Index over specified days. Parameters: days (int): Number of days to analyze (limited by available data). Returns: str: A string containing the analysis results, including latest value, average value, trend direction, and number of data points analyzed.

calculate

Calculate the result of a mathematical expression. Support python math syntax and numpy. > calculate("2 * 3 + 4") {'result': 10} > calculate("sin(pi/2)") {'result': 1.0} > calculate("sqrt(16)") {'result': 4.0} > calculate("np.mean([1, 2, 3])") {'result': 2.0}

calc_ta

Calculate technical indicators using ta-lib-python (TA-lib) and numpy. This tool evaluates a given expression string using the ta-lib-python library. The expression should follow ta-lib-python syntax, for example: - 'ta.SMA(close, timeperiod=30)' with the ticker 'AAPL' - 'ta.ROC(close, timeperiod=30)' with the ticker 'MSFT' - 'ta.RSI(close, timeperiod=14)[-1]' with the ticker 'NVDA' You must specify a ticker to use the cached price data if you've used the 'get_price_data' tool with the same ticker. If not, the expression will be evaluated in a context where only ta-lib-python and numpy are available. The expression string is evaluated in a context where the ta-lib-python library is available as 'ta' and numpy is available as 'np'.

get_current_time

Get the current time in ISO 8601 format.

get_fred_series

Get a FRED series by its ID. However the data is not always the latest, so use with caution!!!

search_fred_series

Search for the most popular FRED series by keyword. Useful for finding key data by name. Like GDP, CPI, etc. However the data is not always the latest.

cnbc_news_feed

Get the latest breaking world news from CNBC, BBC, and SCMP. Useful to have an overview for the day.

README

finance-tools-mcp: A Financial Analysis MCP Server

https://github.com/VoxLink-org/finance-tools-mcp

Overview

The finance-tools-mcp , modified from investor-agent, is a Model Context Protocol (MCP) server that provides comprehensive financial insights and analysis to Large Language Models. It leverages real-time market data, news, and advanced analytics to help users obtain:

Detailed ticker reports including company overview, news, key metrics, performance, dates, analyst recommendations, and upgrades/downgrades.
Options data highlighting high open interest.
Historical price trends for stocks.
Essential financial statements (income, balance sheet, cash flow) formatted in millions USD.
Up-to-date institutional ownership and mutual fund holdings.
Current and historical CNN Fear & Greed Index data and trend analysis.
Prompts related to core investment principles and portfolio construction strategies.
Earnings history and insider trading activity.
Breaking world news from CNBC.
FRED series search results.
Technical indicators using ta-lib-python.

The server integrates with yfinance for market data retrieval and fetches Fear & Greed data from CNN.

It also includes tools for calculating mathematical expressions and technical indicators using ta-lib-python.

And Macro economic indicators using fredapi.

And breaking world news from cnbc.com.

Make sure to also enable web search functionality if you would like to incoporate latest news in your analysis.

Sample Report

alt text

Prerequisites

Python: 3.10 or higher
Package Manager: uv

Installation

First, install uv if you haven't already:

curl -LsSf https://astral.sh/uv/install.sh | sh

Then, you can run the finance-tools-mcp MCP server using uvx:

uvx finance-tools-mcp

If you want to use your own FRED API key, you can set it as an environment variable:

FRED_API_KEY=YOUR_API_KEY uvx finance-tools-mcp

You can also run the server using Server-Sent Events (SSE) transport:

uvx finance-tools-mcp --transport sse

Or with the FRED API key and SSE transport:

FRED_API_KEY=YOUR_API_KEY uvx finance-tools-mcp --transport sse

Usage with MCP Clients

To integrate finance-tools-mcp with an MCP client (for example, Claude Desktop), add the following configuration to your claude_desktop_config.json:

{
  "mcpServers": {
    "investor": {
        "command": "path/to/uvx/command/uvx",
        "args": ["finance-tools-mcp"],
    }
  }
}

Debugging

You can leverage the MCP inspector to debug the server:

npx @modelcontextprotocol/inspector uvx finance-tools-mcp

npx @modelcontextprotocol/inspector uv --directory  ./ run finance-tools-mcp

For log monitoring, check the following directories:

macOS: ~/Library/Logs/Claude/mcp*.log
Windows: %APPDATA%\Claude\logs\mcp*.log

Development

For local development and testing:

Use the MCP inspector as described in the Debugging section.
Test using Claude Desktop with this configuration:

{
  "mcpServers": {
    "investor": {
      "command": "path/to/uv/command/uv",
      "args": ["--directory", "path/to/finance-tools-mcp", "run", "finance-tools-mcp"],
    }
  }
}