Back to MCP Servers

FMP MCP Server

Provider-hosted MCP server from Financial Modeling Prep giving AI agents real-time quotes, fundamentals, filings, and 30+ years of historical market data.

Finance by Financial Modeling Prep API Key active
Overview

The Financial Modeling Prep MCP server is the official remote MCP endpoint maintained by FMP. It wraps the same REST API that powers FMP's developer platform and exposes it as discoverable, schema-defined tools that any MCP client (Claude Desktop, Claude.ai connectors, Cursor, custom agents) can call directly. Once connected, the model can pull stock quotes, company profiles, full financial statements, analyst estimates, SEC filings, insider trades, ETF holdings, macroeconomic indicators, and more without you writing wrapper code.

The server organizes its 250+ underlying tools into roughly two dozen toolsets covering search and directory, company information, financial statements, ratios and key metrics, technical indicators, quotes and historical prices, market indices, news and press releases, SEC filings, insider and institutional trading, ETFs and funds, government (Senate/House) trades, crypto and forex, earnings transcripts and calendars, commodities, and economic data. Inputs are strictly typed (ticker, date range, period), which makes it well suited for agent reasoning over structured financial data.

Authentication is via your FMP API key, appended as a query parameter to the MCP endpoint. The free tier is intended for light testing, and a paid plan is generally required for production use, real-time quotes, and the full historical archive (30+ years of fundamentals).

Tools

Tool Description
search_symbol Search for ticker symbols by company name or partial match.
get_company_profile Retrieve company profile data including sector, industry, CEO, market cap, and description.
get_quote Get real-time stock, ETF, crypto, or forex quote.
get_historical_price Fetch historical OHLCV price data for a symbol over a date range.
get_income_statement Return annual or quarterly income statements.
get_balance_sheet Return annual or quarterly balance sheet statements.
get_cash_flow_statement Return annual or quarterly cash flow statements.
get_key_metrics Retrieve key valuation and performance metrics (P/E, ROE, etc.).
get_financial_ratios Retrieve common financial ratios over time.
get_analyst_estimates Fetch analyst EPS and revenue estimates and consensus price targets.
get_technical_indicator Compute technical indicators like RSI, SMA, EMA, ADX over a price series.
get_market_movers List top gainers, losers, and most active tickers.
get_sec_filings Retrieve SEC filings (10-K, 10-Q, 8-K) for a company.
get_insider_trading Get insider buy and sell transactions for a ticker.
get_institutional_holdings Retrieve 13F institutional holdings for a stock.
get_etf_holdings List underlying holdings and weights for an ETF.
get_earnings_calendar Get upcoming and historical earnings announcements.
get_earnings_transcript Retrieve full text earnings call transcripts.
get_news Fetch stock and general financial news with optional ticker filter.
get_economic_indicator Pull macro data such as GDP, CPI, unemployment, and interest rates.
get_senate_trading Retrieve disclosed US Senate and House trades.
Setup Guide

Prerequisites

  • An FMP account and API key from financialmodelingprep.com. Real-time data and full historical depth require a paid plan.
  • An MCP-compatible client (Claude Desktop, Claude.ai with connectors, Cursor, or a custom agent).

Server endpoint

The official remote endpoint is:

https://financialmodelingprep.com/mcp?apikey=YOUR_FMP_API_KEY

Claude Desktop / Claude.ai (custom connector)

In Claude Desktop or Claude.ai, open Settings, Connectors, Add custom connector and paste your FMP MCP URL as the Remote MCP Server URL:

https://financialmodelingprep.com/mcp?apikey=YOUR_FMP_API_KEY

For Claude Desktop clients that only support stdio servers, bridge to the remote endpoint with mcp-remote:

{
  "mcpServers": {
    "fmp": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://financialmodelingprep.com/mcp?apikey=YOUR_FMP_API_KEY"
      ]
    }
  }
}

Cursor

Open Settings, MCP, Add MCP Server and configure:

  • Name: fmp
  • Type: HTTP
  • URL: https://financialmodelingprep.com/mcp?apikey=YOUR_FMP_API_KEY

Notes

  • Keep your API key out of shared configs; treat the full URL like a secret.
  • Free tier keys are limited in call rate and historical depth. Upgrade for production agents.
Use Cases
  • Fundamental analysis on demand: pull 10 years of income statements, balance sheets, and cash flow for a ticker and have the agent compute margin trends or working capital changes.
  • Earnings prep: fetch the upcoming earnings calendar, prior transcripts, and analyst estimates, then summarize what to watch for an earnings call.
  • Screening and ranking: rank S&P 500 names by ROE, FCF yield, or revenue growth using key metrics and ratios endpoints.
  • Event monitoring: track insider trading, Senate/House trades, and recent 8-K filings on a watchlist.
  • Macro and cross-asset context: combine GDP, CPI, and rates data with equity and crypto quotes for portfolio commentary.
Example Prompts
  • "Pull Apple's last 5 years of income statements and chart revenue and operating margin trends."
  • "Which S&P 500 stocks have ROE above 25% and net debt to EBITDA below 1x?"
  • "Summarize NVIDIA's most recent earnings call transcript and flag any guidance changes."
  • "Show me all insider sales over $1M for TSLA in the past 90 days."
  • "Get the latest CPI and unemployment prints and overlay them with the 10Y yield."
Pros
  • Official, vendor-maintained remote endpoint, no self-hosting or container required.
  • Very broad coverage: 250+ tools spanning equities, ETFs, crypto, forex, commodities, filings, and macro data.
  • 30+ years of historical fundamentals and intraday/real-time quotes on paid tiers.
  • Strict input schemas per tool make it reliable for agentic workflows.
Limitations
  • Real-time data, full historical depth, and useful rate limits require a paid FMP subscription.
  • API key is passed in the URL query string, which is convenient but requires care in how configs are stored and shared.
  • The sheer tool count (250+) can overwhelm smaller models; you may need to scope toolsets per task.
Alternatives