Documentation
QAI Finance
An AI-native research terminal. Read a sector through an interactive research board, drill into a per-company intelligence dossier, ask the QAI analyst a grounded question, and export a presentation-grade report deck — all over web-grounded, primary-source public-market research.
Overview
QAI Finance is a research terminal, not a brokerage. It turns autonomous, web-grounded research over public-market fundamentals into four connected surfaces you can read, interrogate, and present.
- Research boards — a value-chain / thesis map of a whole sector, with positions, shift points, scenarios, and a paper book.
- Company intelligence pages — a deep per-company dossier (financials, margins, moats, ecosystem) that cross-links back into the boards.
- Ask QAI — a research-analyst chat grounded in the board and company data plus live web search.
- Report decks — 16:9 presentation reports you can read on screen or print to PDF.
The terminal lives at qai.io/finance. The home view is the board portal; from there you navigate to any board, company, or report.
Research boards
A board is a structured read of one sector — a value-chain map plus the thesis, positions, and scenarios layered on top of it. Open one from the portal at /finance/boards/[id].
Current boards
| Board | Type | What it maps |
|---|---|---|
| AI Compute Value-Chain | Ecosystem | Margin-pool migration across the AI compute stack under custom-silicon proliferation. |
| Nvidia | Single-name | A focused conviction read on the highest-quality franchise in semis — valuation, segments, moat, catalysts. |
| AI Datacenter Power & Cooling | Ecosystem | The physical ceiling on the AI buildout — grid interconnect, generation, turbines, and liquid cooling. |
| Memory Super-Cycle | Ecosystem | HBM / DRAM / NAND pricing, the cycle second-derivative, the HBM oligopoly, and through-cycle winners. |
| Space / Launch Economics | Ecosystem | The $/kg cost curve, launch cadence and manifest backlog, and the LEO value chain. |
Board sections
An ecosystem board reads top to bottom. Each section is a tab in the board; the single-name board (Nvidia) uses a thesis-shaped variant (Thesis, Valuation, Segments, Moat, Catalysts, Risks, Sizing, Views).
| Section | What it shows |
|---|---|
| Value chain | The layers of the sector, top to bottom, with the names that sit on each — where the margin pools actually live. |
| Shift points | The inflections that move value between layers — what changes, and which names win or lose when it does. |
| Opportunities | The actionable ideas the map surfaces, bucketed by where in the chain they sit. |
| Views (Bull / Bear) + QAI Voices | The two-sided debate on the central call, plus distinct analyst personas arguing it from different seats. |
| $10M book | A paper portfolio that sizes the thesis — positions, conviction, and weight across a notional $10M of NAV. |
| Predictions (4-horizon matrix) | Calls laid out across four time horizons so near-term noise is separated from the structural thesis. |
| Scenarios | Per-name bull / base / bear targets, each with a probability — scenario-conditional, never a point forecast. |
| Regimes | The macro / cycle regimes the thesis is conditioned on, and how the read changes if the regime flips. |
| Premises | The load-bearing assumptions the whole board rests on, each probability-weighted and falsifiable. |
Company intelligence pages
Every covered name has a dossier — a deep, web-grounded read produced by the company-intelligence engine. Browse them at /finance/companies, or open one at /finance/companies/[ticker].
A dossier covers:
- Financials — revenue trajectory, margins, COGS, and capex intensity, each figure stamped with its vintage.
- Earnings — last quarter, beat/miss, guidance, and the key operating metrics.
- SWOT and bull / bear — strengths, weaknesses, opportunities, threats, and the two-sided case.
- Moats, dependencies, bottlenecks — what protects the franchise, what it depends on, and what constrains it.
- Signals and trends — the leading indicators to watch and the structural trends that bend the business.
- Competitors and the ecosystem graph — a value-chain graph of suppliers, customers, partners, and rivals.
How pages cross-link to boards
A company can appear on several boards. Each dossier carries its board context — its role on each board (position, opportunity, book holding, comp, or subject), its side and conviction, and the thesis, catalyst, and falsifier from that board's point of view. So a name reached from a board keeps the board's framing, and a name reached from the directory shows every board it touches.
Ask QAI (the chat)
Ask QAI is a research-analyst chat built into the terminal. It is grounded in the board and company data you are looking at, plus live web search for current facts — and it cites key figures with their vintage.
- Context-aware — open it on a board or a company and it reasons inside that focus; ask from the portal for a cross-coverage view.
- Quantitative and sourced — it prefers a number over an adjective and grounds live facts in web search.
- Two-sided by design — ask "should I buy X" and it answers as analysis: the bull, the bear, and the falsifier — never a bare "buy".
Report decks
Every board generates a presentation-grade report — a 16:9 slide deck rendered in the browser. Open it from a board, or directly at /finance/boards/[id]/report.
| Tier | What it is |
|---|---|
| Executive summary | The default, free landing view — the thesis, the call, and the key scenarios in a tight deck. |
| Institutional full report | The comprehensive deck — the full value chain, shift points, book, scenarios, regimes, premises, methodology, and disclaimers. |
Switch tiers with the Exec / Full toggle in the report header (or the ?tier=full query on the report URL). To export, use Download PDF — it opens your browser's print dialog with each slide laid out one to a page; choose "Save as PDF".
Methodology & data
Boards and dossiers are produced by an autonomous research engine over public-market fundamentals — SEC EDGAR / XBRL filings, 8-K transcripts, and company IR — grounded in live web search for current facts. The method is built to be auditable, not just persuasive.
- Web-grounded, primary-source — figures trace back to primary filings and IR disclosures, not secondary summaries.
- Vintage stamping — every figure carries the period it came from, so you always know how fresh a number is.
- Bull / base / bear with probabilities + falsifiers — targets are scenario-conditional, each with a probability; every position carries an explicit falsifier and kill criteria.
- Four-gate quality pass — each deliverable runs a staged gate: citation-verify, claim-check, red-team review, and a tier-1 audit before it ships to the terminal.
- Company-intelligence engine — the per-company dossiers are generated by a dedicated engine and ingested into the terminal; research publishes, the platform renders.
Geography policy
QAI Finance is US-first. Analysis may reference any company or region where it is relevant to the thesis — a supplier, a competitor, an end market — but the actionable read is US-listed-first.
- Recommendations and sizing are framed around US-listed names.
- A mainland-China name is never framed as a buy or own recommendation — it may appear only as analysis (e.g. a competitor or supply-chain dependency).
Disclaimers
This is not investment advice. QAI Finance is staged analyst work product for a qualified professional — not a recommendation, not a solicitation, not auto-trading.
No order is auto-placed. The terminal does not execute trades, hold custody, or connect to a brokerage. Nothing here moves money.
Bull / base / bear targets are scenario-conditional, not forecasts. Inferred holdings are stamped as inferred. You are responsible for your own decisions; consult a licensed professional before acting.
Use of QAI Finance is governed by our Terms of Service and Fair Use of AI.
FAQ
Is QAI Finance a brokerage or trading product?
No. It is a research terminal. It does not execute trades, hold custody, or place orders. Everything is staged analyst work product for a qualified professional to read and reason with.
Where does the data come from?
Public-market fundamentals — SEC EDGAR / XBRL filings, 8-K transcripts, and company IR — grounded in live web search for current facts. Every figure is stamped with its vintage.
What is the difference between a board and a company page?
A board is a sector-level value-chain and thesis map across many names; a company page is a deep single-company dossier. They cross-link: a board names companies, and each company carries its context from every board it appears on.
What is the difference between the Exec and Full reports?
The executive summary is the default, free landing deck — thesis, call, and key scenarios. The institutional full report is the comprehensive deck with the full value chain, book, scenarios, regimes, premises, and methodology.
Can I export a report?
Yes. Open a report and use Download PDF — it opens your browser print dialog with each 16:9 slide laid out one to a page; save as PDF.
Will Ask QAI tell me what to buy?
No. Ask QAI answers as analysis — the bull, the bear, and the falsifier — never a bare "buy". It is a reasoning tool grounded in the research, not a recommendation engine, and not investment advice.