Financial Datasets MCP and Financial Modeling Prep (FMP) MCP are API passthroughs - strong on fundamentals and SEC filings, but every tool call is a live HTTP request against a rate-limited API. Shibui pre-loads 64 years of data so a single query screens the entire market.
Use Financial Datasets MCP or FMP MCP for deep single-company fundamentals and direct SEC filing access. Use Shibui for cross-market screening, multi-decade backtesting, and research sessions that need to touch thousands of companies in one pass.
Financial Datasets and Financial Modeling Prep (FMP) are two of the most fundamentals-focused MCP servers for Claude. Both connect Claude to financial data APIs via live HTTP requests. Financial Datasets specializes in clean, structured fundamentals and SEC filing extraction from the EDGAR database. FMP exposes 250+ tools covering financial statements, valuation metrics, institutional ownership, earnings calendars, and direct SEC filing access across global markets.
Both are API passthroughs. Every tool call makes a live HTTP request to the provider's servers, and every request counts against your rate limit. FMP's free tier allows 250 requests per day. Financial Datasets offers a free tier with similar constraints. Both require API keys and account registration.
Shibui takes a different approach. Instead of forwarding requests to an external API, it connects Claude to a pre-loaded database with 64 years of US stock data: daily prices, quarterly and annual financials, daily valuations, 56 technical indicators, and 6.4M SEC filing metadata records. A single SQL query can screen the entire market. No API key, no account, no rate limit.
| Capability | Financial Datasets / FMP MCP | Shibui Finance |
|---|---|---|
| Architecture | API passthrough (live HTTP per tool call) | Pre-loaded database (single SQL query) |
| API key required | Yes (both) | No |
| Rate limits | FMP: 250/day free. Financial Datasets: free tier with limits | None - queries run locally |
| Cost | Free tiers with limits; paid plans for higher throughput | Free (requires Claude account) |
| Cross-market screening | One company per request (thousands of calls to scan the market) | Single query screens all ~10,000 companies |
| Fundamental depth | Strong - income statement, balance sheet, cash flow, ratios | Quarterly + annual financials, derived metrics, daily valuations |
| SEC filing access | Direct filing text extraction (Financial Datasets); filing URLs and metadata (FMP) | Filing metadata only (dates, form types, URLs) - no full text |
| Historical depth | Varies (typically 10-30 years depending on data type and plan) | 64 years consistently (1962-present) |
| Technical indicators | Limited (FMP has some; Financial Datasets focuses on fundamentals) | 56 pre-calculated indicators (RSI, MACD, Bollinger, SMA, etc.) |
| Market coverage | FMP: Global (70+ exchanges). Financial Datasets: US-focused | US only (NYSE + NASDAQ) |
| Tool count | FMP: 250+. Financial Datasets: focused set | 11 tools (1 SQL query tool handles everything) |
| Setup | Create account, get API key, configure in Claude | One-click connector, nothing to install |
What are the limitations of Shibui Finance vs Financial Datasets and FMP?
Shibui does not have full-text SEC filing content. Financial Datasets extracts and structures the actual text from EDGAR filings - 10-Ks, 10-Qs, 8-Ks - so Claude can read and analyze the narrative sections, risk factors, and MD&A. Shibui stores only filing metadata (dates, form types, filer names, and URLs). If your workflow involves reading what a company actually said in a filing, you need Financial Datasets or FMP. FMP also covers 70+ global exchanges; Shibui is US-only. Neither Financial Datasets nor FMP are limited to end-of-day data the way Shibui is. The database is also survivor-biased (delisted tickers are removed).
Where the architectural difference matters most
Can Financial Datasets or FMP MCP screen 500 companies in one question?
The fundamental weakness of any API passthrough is cross-market screening. When you ask "find companies with 5 consecutive quarters of revenue growth and P/E below 20", an API passthrough needs to fetch financial statements for each company individually. With FMP's free tier at 250 requests per day, screening 10,000 US stocks would take 40+ days. Even on paid plans, scanning the full market means thousands of sequential HTTP requests.
"Find stocks with revenue growth above 15% for 4 consecutive quarters, positive free cash flow, P/E below 25, and market cap above $2B. Show sector, revenue trend, and operating margin."
Claude writes one SQL query that joins quarterly financials and daily valuations across all ~10,000 companies. The consecutive-quarter check, the cash flow filter, the valuation constraint - all computed in a single pass against data that is already loaded. Results in under a second.
How does full-text SEC filing access compare to filing metadata?
This is where Financial Datasets and FMP genuinely excel. Financial Datasets extracts structured text from SEC filings - you can ask Claude to read the risk factors section of a 10-K, summarize an 8-K, or compare MD&A language across quarters. FMP provides direct links to filing documents and some structured extraction. These are capabilities Shibui does not have.
Shibui stores SEC filing metadata: 6.4M records covering form types (10-K, 10-Q, 8-K, Forms 3/4/5, Schedule 13D/13G, and more), filing dates, filer names, CIK numbers, and EDGAR URLs. You can screen and filter filings at scale - "show all Form 4 insider purchases in the tech sector this month" - but you cannot read the filing text itself through Shibui.
"Find companies where insiders filed Form 4 purchases in the last 30 days and the stock has RSI below 35 and revenue growth above 10%. Show filing date, insider name, and current P/E."
Shibui combines filing metadata with fundamentals, technicals, and valuations in a single query. You cannot read what the filing says, but you can find which companies have recent filings matching specific patterns and cross-reference that against any other data in the database.
Can these API passthroughs combine fundamentals with technicals?
FMP has some technical indicator endpoints, but combining them with fundamentals requires separate API calls - one for the financial statements, another for the indicators, possibly a third for valuation ratios. Each call counts against your rate limit. For a single company this is manageable. For 20 companies across 3 data types, you are looking at 60+ API calls in one conversation.
"Compare the top 5 semiconductor companies by market cap on quarterly revenue growth, operating margin, P/E, EV/EBITDA, and 14-day RSI. Show the last 6 quarters side by side."
Shibui joins financial statements, derived metrics, valuations, and technical indicators in a single query. The data lives in the same database, keyed by symbol and date. No separate tool calls, no rate limit math, no stitching responses together.
What about backtesting with fundamental data?
Backtesting a fundamental strategy requires checking conditions at each historical date across hundreds or thousands of companies. With an API passthrough, this means fetching historical financials for every company at every rebalance date - a combinatorial explosion of API calls. FMP's 250 daily requests would be exhausted checking 25 companies across 10 quarters.
"Backtest: at each quarter-end since 2015, buy stocks with EPS growth above 20%, debt-to-equity below 0.5, and P/E below the sector median. Hold for one quarter. Show average return and win rate by year."
Shibui has consistent quarterly financials and daily prices going back decades. The entire backtest runs as SQL queries against pre-loaded data. No per-company API calls, no rate limits, no multi-day data collection phase before the analysis can begin.
When do fundamentals-focused API passthroughs make more sense?
When your question is about one company and you need depth that goes beyond structured financial data. Reading the actual text of a 10-K filing. Analyzing management commentary in an earnings call transcript. Checking institutional ownership changes from the latest 13F. Accessing real-time fundamentals the moment a filing is published. These are workflows where the API passthrough model - deep, per-company, real-time - is exactly right.
FMP's 250+ tools also cover asset classes and data types Shibui does not touch: global exchanges, forex, crypto, economic calendars, analyst estimates with full historical revisions, and ESG scores. If your research spans beyond US equities or requires data Shibui does not carry, these servers fill the gap.
What Financial Datasets and FMP do better
Full-text SEC filing access is the standout advantage. Financial Datasets extracts and structures the actual content from EDGAR filings so Claude can read, summarize, and compare filing text. This is not metadata - it is the narrative sections, risk factors, and financial tables from the documents themselves. For any workflow that involves reading what a company disclosed, rather than just knowing when they filed, this capability is essential and Shibui does not have it.
FMP's breadth is the other major strength. With 250+ tools, it covers financial statements, valuation metrics, analyst estimates with revision history, institutional ownership, insider transactions, earnings calendars, ESG scores, and more - across 70+ global exchanges. If your research requires a data type or market that Shibui does not cover, FMP is likely the most complete single source available as an MCP server.
Both servers also offer real-time or near-real-time data. Financial statements are available shortly after filing. Price data updates throughout the trading day. For workflows that depend on the latest data the moment it becomes available, rather than waiting for an end-of-day batch update, the API passthrough model delivers data that Shibui will not have until the next morning.
Which tool is right for you?
Use Financial Datasets or FMP MCP when you need to read actual SEC filing text, access real-time fundamentals, research a specific company in deep detail, or work with data types and markets Shibui does not cover (global exchanges, forex, crypto, analyst estimate revisions, ESG, institutional ownership). For single-company deep dives, the API passthrough model works well because you are making a manageable number of calls.
Use Shibui when your questions involve screening across the market, comparing groups of companies, analyzing trends over long time periods, combining fundamentals with technicals in one query, or running research sessions that would exhaust a rate limit. The pre-loaded database means no API calls, no waiting, and no per-company iteration. For a broader comparison of all MCP servers for financial data, see the best MCP server for stock data guide.