Both are free. Yahoo Finance MCP gives you near-real-time quotes from unofficial endpoints that break when Yahoo changes things. Shibui gives you a stable pre-loaded database with 64 years of history and cross-market screening.
Use Yahoo Finance MCP for quick real-time-ish quotes on individual tickers across global markets. Use Shibui for cross-market screening, deep historical analysis, and research sessions that need stable, consistent data.
Yahoo Finance MCP is the most popular free financial data connector for Claude. It wraps yfinance, a Python library that scrapes Yahoo Finance's undocumented public endpoints. No API key, no account, no cost. You get near-real-time quotes, basic financials, and coverage for every ticker Yahoo tracks globally - stocks, ETFs, crypto, indices, mutual funds.
The catch: those endpoints are unofficial. Yahoo has changed or blocked them multiple times. When that happens, the MCP server breaks until yfinance ships a patch. There is no SLA, no support channel, and no guaranteed uptime.
Shibui is also free and also requires no API key. The difference is architectural: instead of scraping external endpoints at query time, it connects Claude to a pre-loaded database with 64 years of US stock data. The data is already there - stable, consistent, and queryable across the entire market in a single pass. The trade-off: US equities only, end-of-day data, no real-time quotes.
| Capability | Yahoo Finance MCP | Shibui Finance |
|---|---|---|
| Architecture | Scrapes unofficial Yahoo endpoints via yfinance | Pre-loaded database (single SQL query) |
| API key required | No | No |
| Cost | Free | Free (requires Claude account) |
| Stability | Unofficial - breaks when Yahoo changes endpoints | Stable - data is local, no external dependencies at query time |
| Cross-market screening | One ticker per request (thousands of calls to screen the market) | Single query screens all ~10,000 companies |
| Historical depth | Varies by ticker (typically 5-10 years daily, some longer) | 64 years consistently (1962-present, all tickers) |
| Real-time quotes | Yes - near-real-time during market hours | No - end-of-day only |
| Market coverage | Global (every exchange Yahoo tracks) | US only (NYSE + NASDAQ) |
| Data breadth | Stocks, ETFs, crypto, indices, mutual funds, forex | US equities and ETFs only |
| Setup | Install yfinance + configure MCP server | One-click connector, nothing to install |
What are the limitations of Shibui Finance vs Yahoo Finance MCP?
Shibui covers US equities only. No international stocks, no crypto, no indices, no mutual funds, no forex. Data is end-of-day - no real-time quotes, no intraday candles, no live price during market hours. If you need to know what a stock is trading at right now, or you need data from the Tokyo or London exchanges, Yahoo Finance MCP handles that and Shibui does not. The database is also survivor-biased (delisted tickers are removed). Yahoo Finance MCP requires installing Python dependencies locally; Shibui's one-click connector has nothing to install.
Where the architectural difference matters most
What happens when Yahoo changes their API?
Yahoo Finance does not publish a public API. The yfinance library reverse-engineers Yahoo's internal endpoints and scrapes them. Yahoo has broken these endpoints multiple times - changing response formats, adding authentication requirements, or blocking automated access entirely. Each time, the yfinance community races to find a workaround, and your MCP server is down until they do.
This is not a theoretical risk. It has happened repeatedly. If your workflow depends on data being available when you ask for it - a Monday morning research session, a daily screening routine, a multi-step analysis you cannot easily restart - the unofficial nature of the endpoints is a real constraint.
"Show me all S&P 500-sized companies where operating margin declined for 3 consecutive quarters but revenue kept growing. Include the margin trend and current P/E."
Shibui's data is pre-loaded and local. There is no external endpoint to break, no dependency on a third party's undocumented infrastructure. The data updates daily after market close, and every query runs against what is already there.
Can Yahoo Finance MCP screen 500 companies in one question?
Like all API-passthrough MCP servers, Yahoo Finance MCP fetches data one ticker at a time. When you ask "find tech companies with revenue growth above 20%", the server has no way to check all companies in one call. It would need to fetch financials for each company individually - thousands of sequential requests to Yahoo's servers. In practice, Claude will check a handful of companies and stop.
"Find stocks with market cap above $5B, revenue growth above 20% YoY, positive free cash flow, and RSI below 40. Show sector, market cap, revenue growth, and P/E."
Claude writes one SQL query that joins financials, valuations, and technical indicators across all ~10,000 companies in the database. The results come back in under a second. No iteration, no per-ticker fetching, no risk of being blocked for too many requests.
How far back does Yahoo Finance MCP go?
Yahoo Finance provides varying amounts of historical data depending on the ticker and data type. Daily prices typically go back 5-10 years for most stocks, sometimes longer for major tickers. Quarterly financials are usually limited to the last 4-8 years. The coverage is inconsistent - different tickers have different start dates, and older data may be missing or incomplete.
"How did stocks with P/E below 10 and positive earnings growth perform over the following 12 months? Show average returns by decade from 1970 to 2020."
Shibui has 31M+ daily price records going back to 1962, with consistent coverage across all tickers for the same time period. Every company in the database has the same depth of history. Multi-decade backtests and cross-era comparisons work because the data is uniform, not stitched together from per-ticker API responses with varying start dates.
Can Yahoo Finance MCP combine fundamentals with technicals in one query?
Yahoo Finance MCP exposes different tools for different data types - one for quotes, one for financials, one for historical prices. To combine fundamentals with technical indicators for a single company, Claude needs multiple tool calls. To do this across many companies means multiplying those calls by the number of tickers. The data arrives in separate responses that Claude has to stitch together in its context window.
"Compare AAPL, MSFT, GOOGL, AMZN, and NVDA on quarterly revenue growth, operating margin, P/E, and 14-day RSI. Show the last 8 quarters side by side."
Shibui joins financial statements, valuations, and technical indicators in a single query. The data already lives in the same database, keyed by symbol and date. Claude writes one SQL statement that pulls from multiple tables and returns a unified result. No separate tool calls, no stitching, no context window overhead from multiple round trips.
When is Yahoo Finance MCP the better choice?
Yahoo Finance MCP wins when you need something Shibui cannot provide: a real-time quote during market hours, a price for a stock on the Tokyo exchange, the current Bitcoin price, or a quick lookup on a mutual fund or index. For these use cases, the unofficial nature of the endpoint is an acceptable trade-off because the query is simple, fast, and does not depend on consistent historical depth.
It is also simpler for single-ticker research where you just want the latest financials and a current price. You ask, it fetches, you get a result. No database to understand, no SQL under the hood. For the question "what is AAPL trading at right now?", Yahoo Finance MCP answers instantly and Shibui cannot.
What Yahoo Finance MCP does better
Zero friction to start. No API key, no account registration, no payment method. Install the Python dependencies, point Claude at it, and you have global market data. For someone who just wants to ask "what is Tesla's P/E ratio?" without any setup ceremony, Yahoo Finance MCP gets there fastest.
Near-real-time quotes during market hours. When the market is open, Yahoo Finance MCP returns prices that are minutes old at most. Shibui updates once after market close. If your workflow involves checking current prices, watching intraday movement, or making decisions based on live data, Yahoo Finance MCP is the only free option that covers this.
Global breadth. Yahoo Finance tracks stocks on virtually every major exchange worldwide, plus cryptocurrency, indices, mutual funds, and forex. Shibui covers NYSE and NASDAQ only. If you research international companies, emerging markets, or non-equity assets, Yahoo Finance MCP is the only free MCP server with that coverage.
Which tool is right for you?
Use Yahoo Finance MCP when you need real-time quotes, international tickers, cryptocurrency prices, mutual fund data, or quick single-company lookups where the unofficial endpoint risk is acceptable. It is the right tool for "what is X trading at?" and "show me Y's latest earnings" - fast, free, global, and good enough for casual research.
Use Shibui when your questions involve screening across the market, comparing multiple companies, analyzing long historical trends, combining fundamentals with technicals, or running multi-step research sessions that depend on consistent data availability. The pre-loaded database means no external calls, no breakage risk, and no per-ticker iteration. For a broader comparison of all MCP servers for financial data, see the best MCP server for stock data guide.