SEC filings
Q1 FY2027 revenue +30%
Signal Bridge reads filings, transcripts, fundamentals, consensus and news for the names you cover, and answers in plain English — with every figure tagged to the source it came from. When the filing doesn't say, it tells you that instead of writing a number.
Free through the beta · founding rate locked at launch · no procurement cycle
Q1 FY2027 revenue +30%
Gross margin 72.85% → 65.20%
Multicloud DB growth 531%
Avg PT $213 vs. $150 spot
Net debt / EBITDA 3.19×
10b5-1 plan, up to 50M shares
The private rule every analyst already has — re-check each number before it goes anywhere — is built into the product instead of into your evening. Source tags are clickable, not decoration, and the model is never the thing that produces a figure.
10-K and 10-Q passages, retrieved and quoted.
Quote, TTM ratios, key metrics, solvency scores.
Four years of income statement, plus latest balance sheet and cash flow.
Analyst estimates, price targets, rating mix.
Industry peers on the same metrics as the subject.
Passages from the latest call, picked against your question.
Company news, with litigation PR filtered out.
Connect a feed your firm already licenses. Your contract, your entitlement — verification applies identically.
Medium-term thesis check
| Factor | Bull case | Bear case |
|---|---|---|
| Revenue ramp | Consensus FY2029E $180B, FY2031E $262B Consensus | $67B base — 4× in five years requires flawless execution |
| Cloud momentum | 121% infra growth; multicloud live across AWS, GCP and Azure Transcript | Gross margin fell 72.85% (FY2023) → 65.20% (FY2026) Financials — capex-heavy infra dilutes mix |
| Free cash flow | Operating CF $32B Financials | Capex −$55.7B → FCF −$23.7B Financials; FCF yield −6.6% Fundamentals |
| Balance sheet | $31.3B cash Financials | $156B gross debt; net debt / EBITDA 3.19× Fundamentals |
| Valuation | P/E 23.17× · EV/EBITDA 14.72× Fundamentals — cheap against peers Peers | Average PT $213 Consensus vs. $150 spot — either the target is stale or the stock is pricing dilution risk |
| Insider signal | Plan is 10b5-1 — pre-scheduled, not reactive News | $7.5B of potential supply; sentiment overhang likely persists |
Ask for something the retrieved context genuinely doesn't contain and the desk declines, names what it does have, and tells you which filing would carry the rest. No plausible-looking number appears to fill the gap.
The context does not contain a quarterly revenue breakdown by segment for the last four quarters. The Financials block provides only annual figures, and the Transcript excerpt covers only Q3 2026 commentary.
What the context does support: annual direction FY2023→FY2026, total revenue $26.97B → $60.92B → $130.50B → $215.94B Financials.
Each agent is a different depth of work on the same verified substrate — a minute, or the better part of a coffee break.
Ask about a company, trade or theme in plain English. Pick which of the seven sources it draws from with the @ menu, or run a named skill instead of a free-form answer.
A multi-agent research pod — specialists, a bear and a synthesizer — argues out a company and lands on a conviction call, grounded in filings. Specialists research concurrently; the bear stress-tests what they found before the call is written.
Everything on file about one company — valuation, financial health, growth, what analysts expect, who owns it — as 27 computed panels, then a 17-section note written in the structure of a research memo.
What the crowd is actually saying, source by source — Reddit, X and a community feed. The composite weights each source by how reliable it has been and how much volume backs it, so three posts don't count like three hundred.
Every company is scored against each lens before a model sees it. When the lenses disagree, the workup says so and gives you the spread rather than averaging the argument away.
Every name you follow is scored against every lens, so you can see at a glance which are worth opening. No AI is involved here — each cell counts how many of that style's tests the company passes, calculated from filed figures. Re-running costs nothing.
| Name | Quality compounder |
Deep value |
Quality growth |
Distress signals ↓ |
Price trend |
|
|---|---|---|---|---|---|---|
| NVDANVIDIA Corporation | 5/5 | 3/6 | 4/5 | 0/5 | 5/5 | ↗ |
| MUMicron Technology, Inc. | 5/5 | 4/6 | 4/5 | 0/5 | 5/5 | ↗ |
| PATHUiPath Inc. | 4/5 | 3/6 | 4/5 | 0/5 | 3/5 | ↗ |
| AMDAdvanced Micro Devices, Inc. | 4/5 | 4/6 | 3/5 | 0/5 | 5/5 | ↗ |
| ADBEAdobe Inc. | 4/5 | 4/6 | 4/5 | 1/5 | 2/5 | ↗ |
| COSTCostco Wholesale Corporation | 4/5 | 3/6 | 3/5 | 0/5 | 1/5 | ↗ |
| ORCLOracle Corporation | 3/5 | 1/6 | 3/5 | 3/5 | 2/5 | ↗ |
Each lens is a set of pass/fail tests calculated from filed figures before any AI is involved. Move a threshold, swap which figure a test reads, hide a lens you don't use, or write one of your own. Changes apply to Deep Analysis and the Screen immediately.
Editing a built-in stores an override, so Reset returns it to the definition in code rather than to your last save. Hiding a lens removes it from the band and the screen without deleting it.
Every symbol is confirmed against live market data before it joins your list, so a typo never becomes a silent gap in a screen. Up to 200 names.
Type /earnings-review ORCL Q3 FY2026 for the desk's full review note on one reported quarter: actuals versus consensus, what changed, the drivers, and what management said. Figures are assembled in code first; the model only writes them up.
Seats two and three see the same coverage list, the same lenses and the same research record. Adding an analyst is a setting, not a second sales conversation.
Every institutional platform in this category is demo-request only, and won't tell you what it costs until you've sat through the call. We'll tell you now — we're just not open to everyone yet.
We're taking a small group of working analysts into the beta. Bring one ticker you know cold and see whether the desk tells you anything you'd have missed — or admits what it can't see.
We read every one and reply within two working days, either way. If it's a fit, the next step is a short call and your coverage list loaded before it.