Fiscal.ai researches individual companies with institutional-grade data. Shibui screens the entire market for multi-year patterns, event outcomes, and every-quarter consistency.
Fiscal.ai (formerly FinChat) is an AI-powered equity research terminal. S&P Market Intelligence data, polished consumer interface, company-specific KPIs like Tesla deliveries or Meta daily active users. 100,000+ companies globally. Plans from free to $99/month. It answers "tell me about AAPL" with depth and precision.
Shibui answers a different kind of question: "find me all companies where X" across the entire US market, with constraints that span 60+ years of history. The distinction is not a feature gap - it is an architectural difference. Fiscal.ai is a research tool for known companies. Shibui is a screening and analysis tool for discovering unknowns.
Different questions, different tools
One company vs. the entire market
Fiscal.ai excels when you already know which company you want to research. "What is Microsoft's free cash flow trend?" or "Show me Tesla's delivery numbers by quarter." It retrieves deep, detailed data about specific tickers with high quality sourcing.
But it cannot answer "find me every company where free cash flow was positive every quarter for 5 years." That requires screening 9,500+ companies simultaneously, checking 20+ quarterly values per company, and returning those that pass. It is a different computational problem.
"Companies where free cash flow was positive every quarter for at least 5 years, with market cap above $5B. Rank by FCF yield."
Claude writes a single query that evaluates every company in the database against the consistency requirement, filters by market cap, and ranks by the derived metric. This is a screening operation across the full universe, not a lookup on a known ticker.
Screening the whole market, not just one company
Fiscal.ai can tell you Apple's profit margin for any quarter you ask about. It can show you the trend on a chart. What it cannot do is screen the entire market for "companies where profit margin exceeded 15% every quarter for 5 years." That is a cross-market consistency requirement - it requires evaluating a condition across multiple periods for thousands of companies simultaneously.
This is not a limitation of Fiscal.ai's data quality (which is excellent). It is a limitation of the single-company research model. The tool is optimized for depth on one ticker, not breadth across all tickers with conditions that must hold across multiple periods.
"Show me companies where return on equity exceeded 20% every year for 10 years, currently trading at a P/E below their own 5-year average."
The consistency check (10 years, every year must pass) combined with a self-referential valuation comparison (current P/E vs. own historical average) requires both market-wide screening and per-company historical computation. Shibui handles both in one pass.
Event studies across the market
Fiscal.ai can tell you what happened to a specific stock after its earnings miss. It cannot answer "across all stocks over 20 years, what was the average price reaction 30 days after an earnings miss greater than 10%?" That requires aggregating outcomes across hundreds or thousands of historical events.
Event studies are statistical research. They answer "what typically happens when this pattern occurs?" by looking at every historical instance, not just one company. This is a fundamentally different question than "what happened to AAPL last quarter?"
"Average stock price change 5, 30, and 90 days after earnings misses greater than 10%, segmented by whether the company had been growing revenue for 4+ consecutive quarters before the miss."
This adds a conditional layer: segment event outcomes by the company's prior trajectory. Are earnings misses more damaging when they break a growth streak? Answering this requires combining sequential pattern detection with event study methodology across the full market history.
Pattern matching on derived time series
Fiscal.ai shows you a company's growth rates, margins, and trends. It does not screen the entire market for companies matching a growth pattern - "revenue growth accelerated for 4 consecutive quarters" or "margin expanded every quarter for 2 years."
Pattern matching requires computing derived metrics (growth rates, ratios) for every company, then checking sequential relationships between consecutive values. It is computationally intensive across 9,500 companies and not something a per-company research tool is designed to do.
"Companies where quarterly revenue growth accelerated for 4 consecutive quarters and operating margin expanded in each of those quarters. Market cap above $2B."
Claude computes sequential growth rates and margin changes for every company in the database, checks both patterns simultaneously, and filters to those where both conditions held over the same four quarters. This is a market-wide pattern scan, not a single-company research query.
What Fiscal.ai does better
Data quality: Fiscal.ai uses S&P Market Intelligence data, which is hand-verified and institutional-grade. Shibui uses tier-3 data providers. For individual company research where precision matters, Fiscal.ai's data is more reliable.
Company-specific KPIs: Tesla vehicle deliveries, Meta daily active users, Netflix subscriber counts, Chipotle same-store sales. Fiscal.ai tracks company-specific operating metrics that do not exist in standard financial databases. Shibui has standard financial statements only.
Standalone interface: Fiscal.ai works in a browser with a polished UI. No Claude subscription required, no MCP connection to configure. The barrier to entry is lower for someone who just wants to research a specific company.
Global coverage: 100,000+ companies across international markets. Shibui covers US equities only - NYSE and NASDAQ, roughly 9,500 symbols.
Which tool is right for you?
If you want to research a specific company in depth - its financials, KPIs, competitive position - Fiscal.ai is excellent. Better data quality, broader coverage, purpose-built interface. It answers "tell me about X" better than Shibui does.
If you want to find companies you do not already know about - screening 9,500 stocks for multi-year patterns, studying what happened historically after specific events, detecting momentum shifts across the market - that is what Shibui does. Different questions, different architecture, complementary rather than competing.