Financial Modeling Prep MCP Server

io.github.imbenrabi由 GitHub 帳號 imbenrabi 發布v2.9.0更新於 Nov 5, 2026

MCP server for Financial Modeling Prep API with 250+ financial data tools

已驗證Streamable HTTP可網頁執行金融資料與分析

概覽

AI 產生的概覽

面向 Financial Modeling Prep API 的 MCP 伺服器,讓助理取得涵蓋股票、加密貨幣、外匯、經濟等領域的 267 個金融資料工具。

可以做什麼
  • 搜尋並瀏覽代號、公司資料與目錄清單
  • 讀取財務報表、指標、比率與分析資料
  • 取得報價、價格資料、技術指標與市場指數
  • 檢索新聞、新聞稿與 SEC 申報文件
  • 查詢內部人、機構與政府交易,以及 ETF 與基金
  • 取得加密貨幣、外匯、大宗商品、財報與經濟資料
適用情境
當助理需要回答上市公司、市場行情、申報文件或總體經濟資料相關問題,且你已有或願意申請 Financial Modeling Prep API 金鑰時適用。託管執行個體適合快速試用,自行架設則適合想自行控制工具載入的團隊。
執行需求
需要 Financial Modeling Prep API 存取權杖,透過 FMP_ACCESS_TOKEN 或 --fmp-token 命令列參數提供。可使用託管的遠端端點,或以 npm/npx 在本機安裝,v2.6.0 以上需 Node.js 25.3.0 或更新版本(v2.5.x 以下需 Node.js 20+)。選用設定:DYNAMIC_TOOL_DISCOVERY、FMP_TOOL_SETS 與 PORT。需能連線至 FMP API 的網路。
安裝前請注意
FMP_ACCESS_TOKEN 是第三方付費 API 憑證;工作階段設定中的權杖可覆寫伺服器本身的權杖,因此不要放入共用連結。查詢內容與權杖會傳送給 Financial Modeling Prep,託管執行個體屬第三方端點。預設模式會一次載入全部 267 個工具,可能佔用大量上下文,動態模式可避免此情況。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Financial Modeling Prep MCP Server,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Web executable,透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式,或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南,先向你展示發布者、端點和登入方式,經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/io.github.imbenrabi%2Ffinancial-modeling-prep-mcp-server/install.md,並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前,先告訴我它的發布者、端點和登入方式,等我確認。不要把 API 金鑰寫進命令或我會提交的檔案;問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "financial-modeling-prep": {
      "type": "http",
      "url": "https://financial-modeling-prep-mcp-server-production.up.railway.app/mcp"
    }
  }
}

README

Financial Modeling Prep MCP Server

[npm version] [license]

A Model Context Protocol (MCP) server for the Financial Modeling Prep API, exposing 267 financial data tools to AI assistants.

Features

  • 267 Financial Tools across 24 toolsets — stocks, ETFs, crypto, forex, commodities, economics, and more
  • Dynamic Tool Management — built on toolception for runtime enable/disable of toolsets via meta-tools
  • Three Server Modes — Dynamic (5 meta-tools), Static (pre-loaded toolsets), or All Tools (default)
  • Flexible Deployment — use the hosted instance or self-host via npm, Docker (built from source), or source
  • Streamable HTTP Transport — compatible with Claude.ai, Claude Desktop, and MCP registries

Quick Start

Hosted Instance (Fastest)

No installation required. Connect directly:

https://financial-modeling-prep-mcp-server-production.up.railway.app/mcp

The hosted instance runs in Dynamic mode: it starts with 5 meta-tools that load toolsets on demand. Pass your FMP API key as FMP_ACCESS_TOKEN in the ?config= session configuration (see Usage).

Self-Hosted (One-Liner)

bash
npx financial-modeling-prep-mcp-server --fmp-token=YOUR_FMP_API_KEY

Or install globally:

bash
npm install -g financial-modeling-prep-mcp-serverfmp-mcp --fmp-token=YOUR_FMP_API_KEY

Get your API key at financialmodelingprep.com.

Table of Contents

Installation

Prerequisites: Node.js 25.3.0 or higher (for v2.6.0+). For older versions (v2.5.x and below), Node.js 20+ is required.

Choose the method that fits your workflow:

  • NPM — npm install -g financial-modeling-prep-mcp-server
  • Docker — build the image from source (no pre-built image is published)
  • Source — clone and run with npm install && npm run build

See docs/INSTALLATION.md for detailed steps per method.

Configuration

The server supports three operational modes controlled via CLI arguments or environment variables:

ModeHow to EnableDescription
DynamicDYNAMIC_TOOL_DISCOVERY=trueStarts with 5 meta-tools; load toolsets at runtime
StaticFMP_TOOL_SETS=search,companyPre-loads specified toolsets on session creation
All Tools (default)(default)Loads all 267 tools immediately

Precedence: for mode and tool sets, CLI args > Environment variables > Defaults. A session ?config= token overrides the server's FMP_ACCESS_TOKEN for that session.

See docs/CONFIGURATION.md for the full configuration reference.

Usage

Connect to the server over streamable HTTP:

  • Claude.ai / Claude Desktop — add as a remote connector (Settings > Connectors)
  • Custom clients — POST to /mcp: send initialize first, then the mcp-session-id it returns on every later request
  • MCP registries — Glama.ai and the official MCP Registry

See docs/USAGE.md for client-specific setup, session headers, and request examples.

Available Tools

24 toolsets covering:

Search · Directory & Symbol Lists · Company Information · Financial Statements · Financial Metrics & Analysis · Technical Indicators · Quotes & Price Data · Market Indexes & Performance · Market Data · News & Press Releases · SEC Filings · Insider & Institutional Trading · ETFs & Funds · Government Trading · Cryptocurrency & Forex · Earnings · Special Data Sets · Commodities · Economics · Bulk Data Tools

See docs/API_REFERENCE.md for the complete tool catalog.

Registries

This server is listed on multiple MCP registries for easy discovery:

See docs/REGISTRIES.md for registry-specific setup instructions.

Contributing

Contributions are welcome. Please see CONTRIBUTING.md for development setup, testing, and pull request guidelines.

License

Apache-2.0

來源:README.md,提交 47dc8f1

工具

5

enable_toolset

Enable a toolset by name

輸入參數name必填

disable_toolset

Disable a toolset by name and remove its tools

輸入參數name必填

list_toolsets

List available toolsets with active status and definitions

無需輸入參數。

describe_toolset

Describe a toolset with definition, active status and tools

輸入參數name必填

list_tools

List currently registered tool names (best effort)

無需輸入參數。

版本歷史

4
  1. v2.9.0最新Nov 5, 2026
  2. v2.8.0Oct 14, 2026
  3. v2.7.0Oct 10, 2026
  4. v2.6.12Sep 16, 2026

評分

已滿足適用檢查項的 100%

該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分