Yahoo Finance MCP Server
Enables Claude to fetch live company data and news by ticker, including quotes, financials, analyst views, and historical data via Yahoo Finance.
README
Yahoo Finance MCP Server
Live company data and news inside Claude, by ticker. Python + official MCP SDK (FastMCP, stdio) + yfinance.
Building a new MCP server for a different site? Start from MCP_PLAYBOOK.md — a self-contained recipe (with the common failure modes) that uses this repo as the reference implementation.
Tools
| Tool | What it returns |
|---|---|
get_quote(tickers) |
Price, day change, day/52w range, volume, market cap — accepts multiple tickers |
get_history(ticker, period, interval, start, end) |
OHLCV bars over a period or date range |
get_financials(ticker, statement, period) |
Condensed income / balance / cashflow, annual or quarterly, + Yahoo link |
get_profile(ticker) |
Sector, industry, description, employees, key stats |
get_valuation(tickers) |
P/E, fwd P/E, EV/EBITDA, EV/Rev, P/B, margins, growth, leverage — multi-ticker for comps |
get_analyst_view(ticker) |
Price targets, consensus, ratings breakdown, EPS/revenue estimates |
get_news(ticker, limit) |
Headlines with source, timestamp, link |
Non-US listings need the exchange suffix: D05.SI (SGX), 0700.HK (HKEX), 7203.T (Tokyo).
Install
cd "D:\Documents\AI Automation\Claude Code Projects\Yahoo Finance"
python -m venv .venv
.venv\Scripts\pip install -r requirements.txt
(Already done — the venv in this folder is ready.)
Register in Claude Desktop
Add to %APPDATA%\Claude\claude_desktop_config.json, then restart Claude Desktop:
{
"mcpServers": {
"yahoo-finance": {
"command": "D:\\Documents\\AI Automation\\Claude Code Projects\\Yahoo Finance\\.venv\\Scripts\\python.exe",
"args": ["D:\\Documents\\AI Automation\\Claude Code Projects\\Yahoo Finance\\server.py"]
}
}
}
Register in Claude Code
claude mcp add yahoo-finance -s user -- "D:\Documents\AI Automation\Claude Code Projects\Yahoo Finance\.venv\Scripts\python.exe" "D:\Documents\AI Automation\Claude Code Projects\Yahoo Finance\server.py"
-s user makes it available in every project.
Register in Claude chat (claude.ai) — custom connector
claude.ai only accepts remote servers, so the server must be hosted at a public
HTTPS URL. The same server.py serves this with the --http flag (Streamable HTTP
at /mcp).
Deploy free on Render (~10 min, one time):
- Push this folder to a GitHub repo (
.venvis gitignored). - At https://render.com → New → Web Service → connect the repo.
render.yamlis picked up automatically (build:pip install -r requirements.txt, start:python server.py --http, free plan). - Copy the service URL, e.g.
https://yahoo-finance-mcp.onrender.com.
Then in any Claude account: Settings → Connectors → Add custom connector →
paste https://<your-service>.onrender.com/mcp → Add (no authentication needed).
Notes:
- Render's free tier sleeps after ~15 min idle; the first tool call after a pause takes ~30–60 s to cold-start, then it's fast.
- The endpoint has no auth — anyone with the URL can query it. It only serves public market data, but don't post the URL anywhere.
- If tool calls fail from the cloud but work locally, Yahoo is rate-limiting the host's IP — redeploy or upgrade the instance.
Run manually (debugging)
.venv\Scripts\python.exe server.py # stdio (Claude Desktop / Code)
.venv\Scripts\python.exe server.py --http # Streamable HTTP on :8000/mcp (claude.ai)
Speaks MCP over stdin/stdout; logs go to stderr. Data is unofficial Yahoo Finance via yfinance — fine for screens and sanity checks, not a substitute for filings or CapIQ on anything that goes in a deliverable.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。