FAQ
How Indian Equity Intelligence works
Short, direct answers. For the full pipeline see the methodology; for every event the engine recognises see the event types.
What is Indian Equity Intelligence (munafa.win)?
Indian Equity Intelligence is a rule-based event-intelligence engine for Indian listed equities. It reads corporate filings published by the NSE, BSE, SEBI, RBI and Indian government bodies, classifies each one, and scores it into an auditable research signal. It is a research dashboard, not a broker, not an investment adviser and not an automated trading system.
Where does the data come from?
Directly from primary sources: the exchanges' own corporate-announcement feeds, SEBI and RBI publications, and ministry and tribunal notices. It does not use press aggregators or secondary reporting, and every signal links back to the filing that caused it.
Does an LLM or AI decide the scores?
No. Every score, action and confidence value is a deterministic function of facts the filing itself states plus market data that has already happened. A language model may, when enabled, phrase an explanation of a fact already extracted and scored, but it can never originate a fact or set a number.
How does a filing become a signal?
The filing is captured byte-exact, classified into one of 104 rule-based event types, and structured facts such as amounts, dates and parties are extracted from its own wording. Materiality (0 to 100) and impact (-100 to +100) are then computed from those facts, checked against what the price has already done, compared with past events of the same type, and finally passed through the NO-TRADE gates.
What does NO-TRADE mean?
Before any action is proposed, a fixed set of hard gates runs, covering things such as an unresolved company, an unverified timestamp, insufficient liquidity and immaterial size. If any one fires, the outcome is NO-TRADE. A screen full of NO-TRADE on a quiet day is the system working: being interesting is not the same as being tradeable.
What happens when a filing does not state a number?
The value stays UNKNOWN. Missing means unknown, never zero: the engine does not fill a gap with a plausible-looking estimate, and the scoring component simply contributes nothing and says so.
Does it place trades automatically?
No. Order placement is reachable from exactly one function in the codebase and requires a single-use confirmation token that expires in five minutes, so nothing can trade without a human confirming it.
How is this different from a news or sentiment feed?
A sentiment feed scores how a story reads. This engine scores what a regulatory filing states, from the primary source, using published rules, and shows the evidence sentence and every gate that was checked. Two runs on the same filing give the same answer.
Which kinds of events does it cover?
104 event types across six categories: corporate actions, financial results and credit, capital structure, management changes, legal and regulatory matters, and macro policy. The full list is on the event types page.
Is this investment advice?
No. It provides automated market research signals. It is not a guarantee of future returns and is not personalised investment advice. Decisions and orders remain the user's own.