Alpha Vantage MCP is an API passthrough - every tool call makes a live HTTP request counting against a rate limit. Shibui connects Claude to a pre-loaded database where a single SQL query screens the entire market.
Use Alpha Vantage MCP for real-time single-ticker lookups across global markets and asset classes. Use Shibui for cross-market screening, deep historical analysis, and multi-step research sessions that would exhaust a rate limit.
Alpha Vantage is one of the most popular financial data APIs, and its MCP server brings that data into Claude. The server exposes 100+ tools covering stocks, forex, crypto, commodities, economic indicators, and news sentiment. Each tool call makes a live HTTP request to Alpha Vantage's API. The free tier allows 25 requests per day. Paid plans start at $49.99/month for 75 requests per minute.
Shibui takes a fundamentally different approach. Instead of forwarding requests to an external API, it connects Claude directly to a pre-loaded database with 64 years of US stock data. When you ask a question, Claude writes a SQL query that runs against 31M+ daily records locally. No HTTP requests, no API key, no rate limit. The trade-off: US equities only, end-of-day data, no real-time quotes.
| Capability | Alpha Vantage MCP | Shibui Finance |
|---|---|---|
| Architecture | API passthrough (live HTTP per tool call) | Pre-loaded database (single SQL query) |
| API key required | Yes | No |
| Rate limits | 25 calls/day (free) or 75/min (paid) | None - queries run locally |
| Cost | Free tier or $49.99+/month | Free (requires Claude account) |
| Cross-market screening | One ticker per call (screening 10,000 stocks = 10,000 calls) | Single query screens all ~10,000 companies |
| Historical depth | 20 years (full history on paid plans) | 64 years (1962-present) |
| Market coverage | Global (US, international, 20+ exchanges) | US only (NYSE + NASDAQ) |
| Real-time / intraday | Yes (1-min, 5-min, 15-min, 30-min, 60-min) | No - end-of-day only |
| Asset classes | Stocks, forex, crypto, commodities, ETFs, options, economic indicators | US equities and ETFs only |
| News and sentiment | Yes - news API with sentiment scoring | No |
| Setup | Get API key, configure in Claude Desktop | One-click connector, no key needed |
What are the limitations of Shibui Finance vs Alpha Vantage MCP?
Shibui covers US equities only. No international stocks, no forex, no crypto, no commodities, no options chains, no economic indicators. Data is end-of-day - no intraday candles, no real-time quotes, no tick data. No news or sentiment analysis. The database is survivor-biased (delisted tickers are removed, which inflates historical returns for small-cap backtests). If your workflow needs real-time prices, global coverage, or non-equity asset classes, Alpha Vantage covers all of that.
Where the architecture difference matters most
Can Alpha Vantage MCP screen 500 companies in one question?
The most common frustration with API-passthrough MCP servers is cross-market screening. When you ask "find companies with revenue growth above 20% and P/E below 15", an API passthrough needs to fetch data for each company individually. With Alpha Vantage's free tier, checking 500 companies takes 20 days. Even on a paid plan (75 requests/minute), scanning the full US market requires thousands of sequential API calls across multiple minutes.
"Find stocks with revenue growth above 20% YoY, P/E below 15, and positive free cash flow in the most recent quarter. Market cap above $1B. Show sector, market cap, and revenue growth."
Claude writes one SQL query that scans all ~10,000 companies in the database. Results come back in under a second. No iteration, no rate limit countdown, no partial results.
What happens when Alpha Vantage MCP hits a rate limit mid-conversation?
A research session that compares 5 companies across multiple dimensions (revenue trend, margins, valuation, technicals) can easily consume 20+ API calls. On the free tier, that is most of your daily allowance in a single conversation. Hit the limit, and Claude either stops or tells you to wait until tomorrow.
"Compare AAPL, MSFT, GOOGL, AMZN, and META on revenue growth, operating margins, free cash flow yield, P/E, and RSI over the past 3 years. Show quarterly trends."
This is a single query against the database. Claude joins financial statements, valuations, and technical indicators in one pass. Five companies, five metrics, twelve quarters - all returned together without burning through a rate limit.
Can Alpha Vantage MCP backtest a strategy across 60 years?
Backtesting requires dense historical data: daily prices, fundamentals at each point in time, and the ability to scan across companies at each historical date. Alpha Vantage provides up to 20 years of daily data per ticker, but retrieving it for hundreds of companies at the volume needed for backtesting would consume thousands of API calls. The data also arrives one ticker at a time, making cross-sectional analysis (ranking all stocks at a given date) impractical.
"Backtest: at the start of each quarter since 2010, buy stocks with RSI below 30, P/E below 12, and revenue growth above 10%. Hold for 90 days. Show average return per cohort and win rate."
Shibui has 31M+ daily price records going back to 1962, with matching technical indicators and quarterly financials. The entire backtest runs as a series of SQL queries against data that is already loaded - no API calls, no waiting, no rate limit math.
What does 100+ tools do to Claude's context window?
Alpha Vantage MCP exposes over 100 tools (one per API endpoint). Every tool's name, description, and parameter schema is loaded into Claude's context at the start of a conversation. This consumes a significant portion of the available context window before you ask your first question, and it makes tool selection harder - Claude has to choose from 100+ options for each step.
"Show me NVDA's quarterly revenue, EPS, and operating margin for the past 3 years alongside its RSI and price performance."
Shibui exposes 11 tools. The primary workflow is three steps: load the schema, load query patterns, then run SQL. Claude's context stays lean, and the single query tool handles everything from simple lookups to complex multi-table joins.
When does the API passthrough model make more sense?
The API passthrough architecture wins when you need data that does not fit in a pre-loaded database: real-time quotes, intraday candles, live forex rates, breaking news with sentiment scores, or economic indicator releases the moment they are published. Alpha Vantage covers all of these across 20+ exchanges worldwide.
If your workflow is "check the current price of AAPL", "get today's EUR/USD rate", or "show me the latest Fed funds rate", the rate limit is irrelevant - you are making one or two calls, not thousands. Single-ticker research that needs real-time or global data is exactly what the API passthrough model is designed for.
What Alpha Vantage MCP does better
Alpha Vantage covers the global market. US exchanges, London, Tokyo, Hong Kong, Frankfurt, Toronto, Sydney - more than 20 exchanges across the world's major financial centers. If you are researching international equities or comparing companies across multiple markets, Shibui cannot do that. It is US-only.
Asset class breadth is the second major advantage. Forex pairs, cryptocurrency prices, commodity futures, options chains, economic indicators (GDP, CPI, unemployment, Fed funds rate) - Alpha Vantage covers all of these through a single API. Shibui only handles US equities and ETFs. If your analysis spans multiple asset classes, Alpha Vantage is the only MCP server that covers them all in one place.
Brand maturity and ecosystem size also matter. Alpha Vantage has been operating since 2017, has well-documented APIs, extensive community tutorials, and integrations with dozens of platforms beyond MCP. The API is battle-tested across millions of developers. Shibui launched in 2025 and covers a narrower slice of the financial data world.
Which tool is right for you?
Use Alpha Vantage MCP when you need real-time quotes, intraday data, international stocks, forex, crypto, commodities, options, economic indicators, or news sentiment. For single-ticker research across these dimensions, the rate limit is a non-issue and the global coverage is unmatched. If you are on a paid plan and your workflow stays within 75 requests per minute, Alpha Vantage handles most single-company research comfortably.
Use Shibui when your questions involve screening across the market, comparing multiple companies, analyzing long historical trends, or running multi-step research sessions that would burn through rate limits. The pre-loaded database means no API calls, no waiting, and no limits on how many companies or time periods you include in a single question. For a broader comparison of all MCP servers for financial data, see the best MCP server for stock data guide.