Following along? Open the live page
Finmagine — free to explore • premium for full access • no app needed
Five research layers — fundamentals through concall analysis — assembled by Claude in one prompt
An equity research analyst building a comprehensive note on a company would typically spend 2–3 hours sourcing data from SEBI filings, BSE announcements, concall transcripts, financial databases, and management disclosures. The Finmagine MCP deep dive condenses that into a single Claude prompt that calls 15–20 tools in sequence and synthesises the findings into a structured investment case.
The workflow is organised in five layers — from the outermost (market momentum) to the innermost (what management actually said on the last concall). Each layer adds a new dimension to the investment picture.
Each layer moves from the quantitative to the qualitative. You cannot interpret a management tone score without first knowing whether the company's fundamentals are sound — which is why the workflow runs in a fixed sequence.
get_company_profile is the entry point. It returns the full snapshot: PE, PB, ROCE, ROE, D/E, OPM%, 5-year revenue and PAT CAGRs, FCF yield, pledge %, RS rating, 52-week range, and EMA levels. Read it as a health certificate before going deeper.
get_ratios adds the extended set including the Altman Z-score, OCF/PAT ratio, and profitable years over 10 years. The red flags that should pause any analysis:
| Metric | Red Flag Threshold | What It Signals |
|---|---|---|
| Altman Z-score | < 1.8 | Financial distress probability — check debt and cash burns |
| Pledge % | > 25% of promoter holding | Forced-selling risk if price falls; above 50% = serious concern |
| D/E ratio | > 1 (non-financial) | Leverage amplifies downside; check if debt is deployed productively |
| OCF/PAT | < 0.5 | Earnings not backed by cash — potential accounting quality concern |
| Profitable years 10Y | < 7 | History of losses means the current profitability may not be durable |
get_company_scorecard returns a pre-computed 5-dimension rating: financial health, growth prospects, competitive position, management quality, and valuation. Use it as a comparative benchmark — is this company in the top quartile on management quality within its sector, or the bottom half? Omit the symbol parameter to get a full ranked list sorted by overall score.
get_financials with period=quarterly&limit=8 returns 8 quarters of sales, EBITDA, PAT, and OPM%. The key question: is revenue growth accelerating, stable, or decelerating? Is OPM% expanding (operating leverage in action) or compressing (cost pressures)?
The most important thing to establish from 8 quarters: Is the trend improving or deteriorating? A company with declining OPM% for 4 straight quarters that is "still profitable" is a very different investment from one with consistent 200bps annual margin expansion.
get_momentum returns RS rating, the 30-day RS history, and four boolean signals: above EMA-10, above EMA-21, EMA-10 above EMA-21, and near 52-week high (price ≥ 95% of 52W high). Improving RS history + above both EMAs is a confirming technical signal that the market is agreeing with the fundamental case.
get_ecs_score returns the ECS composite, the company's rank within the Nifty 500 universe for the latest quarter, and the PEAD signal. The ECS measures seven dimensions of earnings quality: Net Profit growth YoY, Net Profit growth QoQ, Sales growth YoY, Sales growth QoQ, EBIT growth YoY, EBIT growth QoQ, and CFO/PAT ratio. A top-quartile ECS rank (250th or better in the Nifty 500) means the company delivered strong results on most or all dimensions.
get_company_pead shows whether this specific company has an active post-earnings drift event. A positive PEAD signal means the stock showed an above-average volume surge at earnings, and price has continued to drift in the same direction for 5–45 days. This is the Post-Earnings Announcement Drift anomaly in real time — markets underreacting to strong results.
get_ai_quarterly_summary returns a structured AI-generated earnings brief with a headline verdict, key insights, and the most important data points from the quarter. It covers 703 companies for Mar 2026. For any company with an AI summary available, this is the fastest way to understand what happened in narrative terms — not just numbers.
This is where the deep dive separates from a standard data pull. Finmagine processes concall transcripts from ~400 companies and extracts three layers of management signal.
get_guidance_tone returns a sentiment score (-1 to +1), a tone label (positive/neutral/negative), and counts of guidance items that beat, missed, or came in line with prior guidance. Use this to answer: Is management confident or hedging?
get_management_tone adds the guidance_score (0–100), tone trajectory across quarters, phrase drift signals, and investor lens. The tone trajectory is particularly valuable — a company whose guidance tone has been declining for 3 straight quarters before a sudden positive quarter may be doing PR, not delivering real improvement.
get_concall_decoder is the most granular management analysis tool. It tracks:
This is the tool that surfaces management credibility issues that do not appear in any financial statement. A management team that consistently promises capex deployment and consistently delays it is a different risk than the numbers suggest.
get_ai_company_profile returns the full 10-section AI-generated profile: business model, competitive moat, management quality, growth drivers, risk factors, financial health, industry dynamics, recent developments, investment thesis, and explicit bull and bear cases. Available for ~400 companies.
get_shareholding returns 8 quarters of promoter/FII/DII/public holding. Rising FII holding in a stock with improving fundamentals is confirmation that informed institutional investors are positioning. Falling promoter holding over multiple quarters (not a one-time sale) warrants investigation.
get_pledge_history gives 12 quarters of pledge % as a share of promoter holding. A declining pledge trend — promoters de-pledging their shares — is a bullish signal (they are retiring the loans, reducing forced-selling risk). A rising trend is the opposite.
get_insider_trades with type=buy&days=180 shows promoter and director purchases in the last 6 months. Cluster buying — multiple insiders buying in a short window — is a higher-conviction signal than a single transaction.
get_order_wins returns the company's contract win history with deal value, client, and segment. For capital goods, infrastructure, and defence companies, the order book is a leading indicator of revenue 12–24 months out. A company with a growing order book but declining current revenue is a different situation from one with both declining.
Replace TARGET_SYMBOL with your NSE ticker (e.g., RELIANCE, HDFCBANK, LT). The prompt runs all 5 layers sequentially and ends with a structured Buy/Hold/Avoid verdict.
Add get_banking_kpis in Layer 5 and make it the primary financial health lens. Standard ROCE and OPM% are not meaningful for banks — NIM, GNPA/NNPA, CASA ratio, and credit growth are the metrics that matter. A bank with a declining GNPA trend and stable NIM is fundamentally improving even if its PE looks high.
Heavily weight get_order_wins and the order book growth rate. For these companies, the P&L is a lagging indicator — revenue today is the result of orders won 12–24 months ago. A company with a 2x order book expansion that is not yet in revenue is a very different situation from what its trailing financials suggest.
Always run get_conglomerate first for Tata, Adani, Bajaj, or Mahindra group companies. The parent group's financial health (especially leverage levels), news overhang, or sector rotation can dominate individual subsidiary performance regardless of the subsidiary's own fundamentals.
For ~150 companies in ~25 specific sectors, get_sector_kpis returns operational metrics from investor presentations — for cement: capacity, volume, realization per tonne, EBITDA/tonne; for power: PLF, capacity utilisation, tariff realisation. These granular KPIs are the real drivers that standard P&L ratios obscure.
The Finmagine MCP server includes 24 us_* tools for NYSE/NASDAQ research. The US deep dive covers the same five research layers — adapted for US market data structures: USD financials, Form 4 insider trades instead of NSE disclosures, SEC 8-K filings instead of BSE announcements, and 13F superinvestor holdings instead of DII/FII data.
Finmagine gives you 30+ computed financial ratios, sector benchmarks, FII/DII flows, the Finmagine Score, and AI-powered analysis — all in one place.