Evidence-grounded equity research
Research you can audit, line by line.
A deterministic engine sets every signal from evidence it can trace to a provider. Disagreements and gaps stay visible. An explanation layer then narrates the result — and has to cite the evidence for every sentence it keeps.
Free: 5 research runs a day, no card required.
Largest contributions to the composite
- 63-day momentum+0.060
- Relative strength vs SPY+0.058
- 52-week-high proximity+0.054
- Volume confirmation-0.042
- Gross profitability (GP/A)+0.031
How a run is assembled
The signal is fixed before a word is written
Four stages, always in this order. The explanation layer reads the engine’s output; nothing it writes can change it.
- Providers answeror say that they did not. A missing input stays missing.
- Evidence is reconciledreadings are compared across vendors, never averaged into one.
- The engine sets the signaldeterministic factor math, a macro gate on momentum, itemised confidence.
- The narrative citesand is validated against the evidence. Failing sections are withheld.
The company workspace
One company, every layer of the evidence
These are the terminal’s own panels, rendered from one recorded production run — not a mock-up.
The verdict, decomposed
Five factor families, their weights, and every factor the engine computed — its value, its standardized score and what it contributed.
Every factor
value, standardized score and contribution to the composite| Factor | Family | Value | z | Contribution | |
|---|---|---|---|---|---|
| 63-day momentum | momentum | 0.1509 | 1.66 | +0.060 | |
| Relative strength vs SPY | momentum | 0.0759 | 1.58 | +0.058 | |
| 52-week-high proximity | momentum | 0.9919 | 1.42 | +0.054 | |
| Volume confirmation | momentum | 0.8079 | -1.05 | -0.042 | |
| Gross profitability (GP/A) | quality | 0.5434 | 1.42 | +0.031 | |
| 21-day momentum (timing) | momentum | 0.0860 | 1.63 | +0.030 | |
| Earnings yield | fundamental | 0.0257 | -0.97 | -0.023 | |
| Asset growth | quality | -0.0157 | 0.64 | +0.015 | |
| P/E gap to sector | fundamental | 0.0889 | 0.59 | +0.014 | |
| Net share issuance | quality | -0.0227 | 0.45 | +0.011 | |
| News sentiment | news | 0.0637 | — | +0.010 | |
| Short-term reversal | reversal | 0.0139 | -0.37 | -0.009 | |
| Analyst target upside | fundamental | -0.0261 | -0.16 | -0.004 | |
| Post-earnings drift | fundamental | 6.7400 | 1.35 | +0.003 |
Signal decomposition
composite +0.21| Family | Weight | Score | Direction |
|---|---|---|---|
| Momentum | 40% | +0.40 | |
| Fundamental | 20% | -0.05 | |
| Quality | 15% | +0.38 | |
| News | 20% | +0.05 | |
| Reversal | 5% | -0.18 |
Macro gate ×0.99 applies to the momentum sleeve only · stress probability 1.7%
Largest contributions
- +0.06063-day momentummomentum
- +0.058Relative strength vs SPYmomentum
- +0.05452-week-high proximitymomentum
- -0.042Volume confirmationmomentum
- +0.031Gross profitability (GP/A)quality
Recorded production run · AAPL · 24 Sep 2026, 12:37 UTC. Rendered by the terminal’s own components.
Methodology · scoring-v2.1
Five families, one composite, a gated sleeve
Each factor is normalized with outlier-resistant statistics against the company’s own history and volatility, then weighted into its family; the families are weighted into one composite. Macro stress scales only the momentum sleeve — value, quality and news are never macro-suppressed — and in high-volatility regimes momentum is halved in favor of reversal.
Trend persistence across one to twelve months, confirmed by volume and measured against the market.
63-day momentum · 21-day momentum (timing) · Volume confirmation · 52-week-high proximity · Relative strength vs SPY
What the price pays for earnings, against the sector and the street, and the drift after results.
Analyst target upside · Earnings yield · P/E gap to sector · Post-earnings drift
How profitably assets are used, and whether the share count and balance sheet are being diluted.
Gross profitability (GP/A) · Net share issuance · Asset growth
The tone of recent company headlines, weighted up around earnings.
News sentiment
Short-term overextension that tends to mean-revert, as a small contrarian counterweight.
Short-term reversal
Weights and factors as used in the recorded run. The momentum gate there was ×0.99.
Confidence and risk, itemised
Recorded production run · AAPL · 24 Sep 2026, 12:37 UTCConfidence composition
| Component | Points |
|---|---|
| Model confidence base | 100 |
| Less: Family dispersion | −21 |
| Less: Data completeness | −12 |
| Less: Model reliability (unmeasured) | −10 |
| Less: Data freshness | −3 |
| Less: Macro uncertainty | −1 |
| Less: Conflicting signals | −1 |
| Confidence | 53 |
Risk composition
| Component | Percentile | Weight | Contribution |
|---|---|---|---|
| tail risk | 66.4 | 0.15 | 9.96 |
| idiosyncratic | 94.6 | 0.10 | 9.46 |
| downside dev | 34.6 | 0.20 | 6.91 |
| vol regime | 41.7 | 0.13 | 5.43 |
| beta | 52.2 | 0.10 | 5.22 |
| liquidity | 39.6 | 0.08 | 3.17 |
| drawdown | 25.9 | 0.12 | 3.11 |
| sector | 50.0 | 0.05 | 2.50 |
| macro | 1.7 | 0.07 | 0.12 |
| Risk score | 46 |
Evidence quality strong — The evidence behind this analysis is complete, fresh and internally consistent.
- Unknown is not zero. A factor that could not be computed is reported as missing and costs confidence.
- One source is not agreement. Readings from a single vendor are labeled single-source.
- Disagreement is shown, not averaged. Conflicting vendor figures appear side by side.
- Stale is not current. Every observation carries its own date.
Latest
Market news, with its sources
Market news
aggregated from public financial feedsPricing
Free to use daily. Pro when you need depth.
Free
₹0
- 5 research runs a day
- Full signal, evidence record and synthesis
- Three-month price charts
- Macro conditions from FRED
- Headlines, without article links
Pro
₹100one-time
- Unlimited research runs
- Every chart window, one month to five years
- Full article access from the news evidence
- Everything in Free
Payments through Razorpay. One payment, no renewal.
Questions
Does an AI decide the signal?
No. The signal, confidence and risk level come from a deterministic engine — the same inputs always give the same output. A language model writes the explanation afterwards; it must cite the evidence items it uses, and any section that cites nothing, or quotes a number the evidence does not contain, is withheld. The page says which sections were withheld and which provider and model wrote the rest.
Where does the data come from?
Prices, fundamentals and news come from the market-data vendors configured for the deployment; filings come from SEC EDGAR and macro series from FRED. Every run lists which providers answered, which failed and which inputs are missing. A missing input is never filled with an estimate, and a single source is never presented as agreement.
Is this investment advice?
No. OmniSignal is a research and education tool. It summarizes public data into a structured, repeatable readout; it does not know your situation, and a signal is a summary of evidence, not a recommendation. Decisions and their consequences remain yours.
Is Pro a subscription?
No. Pro is a single ₹100 payment through Razorpay that unlocks the paid features on your account. There is no recurring charge and nothing to cancel.