
Siddhi Nagvekar
Cofounder @ FiDUS
BITS PilaniB.E. (Hons.) Electronics and Electrical, with M.Sc. Economics
FRM Part I cleared
Our flagship
FiDUS Desk
FiDUS Desk is an agentic research system built on every Indian company's filings. It sources ideas from value chains and screens, builds the analysis, holds your written thesis against each new filing and call, and tells a portfolio manager what changed on the book. Brokers use it to write and refresh coverage in their own format. It produces evidence. The decision, and every regulated call, stays with you.
01/For funds, buy side
Research that checks itself.
FiDUS finds the idea, builds the analysis, then holds your written thesis against every new filing, result and earnings call and tells you which reason changed, and why. For a portfolio manager it does the same across the whole book. You get the evidence, never a buy or sell call.
Funds read the market's claims against the filings. Brokers write them from the filings. Same evidence, same rules, two seats.
For funds, buy side
Check it in detailIdeas sourced from structure, not from a feed.
Value chain · Indian pharmaceuticals · 1 of 62
Verified economics × choke × coverage gap
Median EBITDA margin of verified companies in each node. The live screen also weighs the part of the pocket your book is missing.
The analysis, built from the filing itself.
Statement line
Revenue from operations
27,958.0
LUPIN · FY2026 · ₹ crore
LUPIN · FY2026 · consolidated · income statement · ₹ crore
| Revenue from operations | 27,958.0 |
| Other income | 424.5 |
| Income | 28,382.5 |
| Cost of materials consumed | 4,621.9 |
| Employee benefit expense | 4,574.5 |
| Other expenses | 7,241.0 |
| Profit before tax | 6,872.6 |
| Profit for the period | 5,355.5 |
13 of 44 filed lines · filed 7 May 2026
nsearchives.nseindia.com/…/INTEGRATED_FILING_INDAS_1664747_07052026113234_WEB.xml
143 checks · 9 frameworks · the question each one asks
Is the growth, margin and cash story what the P&L and balance sheet actually show?
Do the accounts show the usual signs of aggressive or unreliable reporting?
Are reported profits backed by cash, or flattered by working capital and accruals?
Do the people running it act like owners, and say bad news as plainly as good?
Is the thesis resting on a story, or would it survive an outside view?
Is it a durable, high-return business, and what protects it?
Is the industry over-building or over-lending ahead of a downturn?
What does the price already assume, and how much room is left if wrong?
Does each rupee retained or returned create value per share?
One check, opened up · Earnings quality · 1 of 15
runs as a calculation
Receivables and inventory vs sales growth
What it checks
Why it matters
Red flag
Computed as
Evidence read
Then the analyst
A thesis that stays honest.
EBITDA margin reaches about 25% in FY27.
Revenue grows high single digit.
No generic entry on Tolvaptan or Mirabegron before FY28.
Pithampur remediation closes without an import alert.
bseindia.com/xml-data/corpfiling/AttachLive/5b9e63ec-7102-47fa-b4c9-37b33c391a83.pdf
A directional call scoring under 55 is demoted to needs review: 52 is under 55. The other two assumptions pass through untouched.
The book, not just the stock.
The book · illustrative, not live
Alerts, ranked by position size
One question, many agents.
Ask anything about your coverage or your book. A supervisor routes the question to specialist agents (financials, calls and promises, forensics, valuation, peers and value chains, portfolio) and returns one answer in which every claim carries its source.
BuildingFor brokers, sell side
Check it in detailCoverage that refreshes itself.
Results day · Q1 FY27 filed
₹0 cr
Revenue cited · filing
0.0%
EBITDA margin cited · filing
The rating, the target price, the recommendation, the disclaimer and the contact block are your analyst's alone. The page reads NOT RATED until your analyst signs.
Every other tool in this market rents its numbers from a vendor that forces each company into the same template and loses whatever does not fit. We parsed 2,510 companies as they were filed, so a company's own line items survive, every figure traces to its filing and page, and a backtest cannot see the future. The Desk runs on it. You can run on it too.
HDFC Bank, 1.01x leverage.
A US-schema vendor reads about 6.3x because it counts customer deposits as debt. We report them separately and exclude them.
Reliance, revenue by segment.
The segment table is tied to the figure Reliance filed in XBRL, and the proof comes back with the answer. A table that cannot be tied is marked unverified.
Point in time.
Ask as of 31 March 2024 and only filings filed by that date are considered. No look-ahead.
Four things hold true on both seats of the Desk and in the data layer: where a number came from, whether it could have been known on a given date, what happens when data is missing, and who makes the call at the end.
Read the filing as the company actually submitted it (statement, note, call or deck), not a vendor's normalized restatement.
Hold the source page on every figure, so a number can be walked back to the exact filing it came from.
Answer historical questions using only what had been filed by that date. Missing data is withheld and counted, never assumed.
Hand the evidence to the analyst or the agent working alongside them. The judgment, and every regulated call, stays theirs.
No stock picks.
Every product produces evidence, never a buy or sell call.
No guaranteed returns.
Every figure is a measurement on a stated date, and markets move.
The analyst stays in control.
Nothing regulated (a rating, a target price, a recommendation) is machine-written.
A competitor can copy the data in a quarter. What it cannot copy is what only exists because we ran. Every thesis, guidance statement and check in the Desk is a claim, the evidence behind it, and how the filings eventually settled it. We keep that record. Three parts of it cannot be rebuilt afterwards, because the moment to capture them is the day they happen.
What management promised, and what they filed.
Each guidance statement is held open until a filing settles it, building a per-company record of hit rate and miss size.
What each thesis said would break it.
Falsifiers set today are settled by the filings of the quarters after.
What we believed, and when.
Every restated figure keeps both versions with the date we learned it, from 17 July 2026.
The desks that put FiDUS evidence in front of their own decisions.
FiDUS data layer · figures measured 2 October 2026
The data was wrong in places nobody had looked. These are three we caught while parsing the filings ourselves.
A 1:5 split read as an 80% crash.
NSE's previous-close field is adjusted for only 2 of 549 split ex-dates, so anyone who trusts it sees a collapse that never happened. We adjust from the corporate action itself.
A general insurer with a 17.2% return on assets.
One form nets policyholder liabilities before it totals, so its assets tag is the application of funds. ICICI Lombard tags 16,072 crore against real assets of 75,994 crore. We rebuild total assets from the form's components, and the true figure is 3.6%.
One in six transcripts was a cover letter.
1,711 of 11,205 documents were one or two pages pointing at the company's website, so a search said revenue appeared nowhere in the call. We recovered the real transcripts and flag any that remain.
Everything on this page carries a tag. Here is the same list in one place, so you can see where the Desk is today.
Live now
Building
Tell us who you are and which seat you want to see. A person replies, not a sequence. If you run a fund or a research desk and want to put the Desk on live work, say so: we are looking for design partners.
FiDUS is built and run by two people. Write or call either of us directly.

Cofounder @ FiDUS
BITS PilaniB.E. (Hons.) Electronics and Electrical, with M.Sc. Economics
FRM Part I cleared

Cofounder @ FiDUS
BITS PilaniB.E. (Hons.) Electronics and Instrumentation
CFA Level I cleared · FRM Part I cleared