About us

Infrastructure that turns information into decisions.

Quantify Terminal ingests public and private information, compiles it into structured financial knowledge, propagates change through causal relationships into live models and machine-readable theses, recomputes portfolio-level consequences, and presents decisions inside explicit mandates and human approval gates. All of it runs behind one institutional desktop terminal.

Explore the Platform Institutional Enquiries

Terminals tell investors what happened. Quantify continuously determines what changed in the investor's evidence, assumptions, models, risk and decisions because it happened.

Search, summarisation and generic AI research are baseline capabilities now. They do not maintain an investor's financial state. That state lives in spreadsheets, notes, models and memory, and it goes stale the moment the world moves.

Quantify is built as an event-to-decision system. Filings, earnings, news, macro prints and private documents enter the Data Fabric with canonical identifiers, point-in-time timestamps and machine-readable data rights. The Knowledge Compiler turns them into facts, claims and evidence. The Causal Market Engine traces the change through suppliers, customers, geographies and factors; the Expectation Engine measures it against consensus and what price already embeds. The Financial Model Compiler proposes cell-level updates with source lineage. The thesis graph re-scores every assumption. The Portfolio Digital Twin recomputes first-, second- and third-order exposure. The Decision Engine tests the result against the investor's constitution and writes the outcome to the Decision Ledger.

Every transition stores provenance, point-in-time state, permissions, model version and an audit event. Language models plan, interpret and explain. Financial mathematics runs in deterministic, versioned code.

All of it surfaces inside a native desktop terminal that covers live markets, research, portfolios, risk, derivatives, funds, macro and geopolitics on macOS, Windows and Linux.

We build the system that maintains conviction, not just information.

Architecture

Six layers, one system.

No disconnected tabs. Every layer shares the same facts, rights, evidence graph, model registry, thesis graph and Portfolio Twin, so a change at the bottom reaches a decision at the top with its provenance intact.

P0 · Foundation Data Fabric and World Model

Canonical identifiers, a universal field namespace, point-in-time observations, machine-readable data rights and a temporal entity graph across companies, securities, people, facilities, suppliers, customers, funds, countries, regulators and commodities. Every downstream system shares the same facts and the same permissions.

P1 · Intelligence Pulse, Memory and Deep Research

An always-on anomaly and relevance engine; durable institutional memory of every session, valuation snapshot, thesis, decision and outcome; and a multi-step evidence engine with deterministic financial tools, a verifier, narrative clustering and source reliability scoring.

P2 · Action Workflow Engine, Decision Engine and Scenario Lab

Natural-language objectives become explicit plans with permission classes and an execution trace. Mandates compile into an executable Investment Constitution. Forward shocks, macro moves and counterfactuals run with visible causal paths and model sensitivities.

P3 · Network Investor Chat

Rooms attached to tickers, themes, private deals, real estate and investment clubs, each carrying live financial objects. Ticket-size gated data rooms enforced server-side, research battles, timestamped track records, assumption distributions and network intelligence.

P4 · Institutional Event, forensic and private-market systems

Live event intelligence for earnings and central-bank decisions with expectation gaps and guidance parsing; forensic accounting and an executive truth engine; an alternative data studio with lineage and leakage control; an autonomous diligence room; a market recorder.

P5 · Decision systems Model Compiler, thesis graph, Portfolio Twin, Decision Ledger

Filings compile into versioned models with cell-level lineage. Theses carry dependencies, health and invalidation. The Portfolio Digital Twin runs failure search, hedge design and liquidity stress. Research runs through Financial Git and CI/CD, and every decision is recorded and replayable.

Investment Constitution · executable mandate

max_single_position          = 0.07
max_effective_china_exposure = 0.25
min_expected_return          = 0.15
require_thesis               = true
require_bull_base_bear       = true
reject_if_accounting_risk    > 70
reject_if_liquidity_days_to_exit > 5

How we build

Engineering principles.

The rules the system runs on. They are constraints in the architecture, not statements of intent.

Deterministic finance

Language models plan, interpret and explain. DCF, WACC, ROIC, factor, risk and backtest mathematics run in versioned code, never in prose.

Point-in-time by default

Every observation carries effective, published, available and ingested timestamps. Research and backtests use what was knowable at the time, not what was restated later.

Data rights as infrastructure

Each source carries a machine-readable rights profile for display, calculation, caching, redistribution and AI retrieval, enforced at the API gateway and in every downstream job.

No magic scores

Thesis health, Pulse scores and backtest integrity scores always expose their components, inputs, uncertainty and evidence. A number without a breakdown does not ship.

Human gates on capital

Portfolio changes require explicit confirmation and an audit record. Live orders require hard confirmation, broker controls and a kill switch. Nothing trades by accident.

Permission-aware intelligence

Embedding access is data access. Vector indexes preserve document ACLs, room tiers and client isolation, so restricted material never surfaces through an AI answer.

Reproducible research

Every report stores dataset versions, source IDs, query plan, calculation, model and prompt versions and permissions. A reviewer presses Reproduce and gets the same analysis from the same snapshot.

Evidence or nothing

Every agent argument cites evidence and deterministic calculations. Bull, bear, forensic, macro, industry, quant and risk perspectives use different evidence paths, and their disagreement stays visible.

Capabilities

What the platform covers.

The desk layer and the intelligence layer, from the first screen to the recorded decision.

Live market data Equity research Valuation & DCF Portfolio analytics Risk & VaR Options & derivatives Funds & income Screeners & heatmaps Quant Studio & backtesting Algorithmic execution Broker connectivity Macro & geopolitics Supply chain graph Spreadsheet API Pulse Deep Research Financial Model Compiler Thesis Health Portfolio Digital Twin Scenario Lab Decision Ledger Investor Chat Agent Runtime & API

Company

Leadership and direction.

Quantify Terminal is led by founder and Chief Executive Officer Aaryan Saroha, who directs architecture and product across the Data Fabric, the finance engine, the desktop terminal and the systems that connect them.

Meet the Founder Careers

Get started

Evaluate the terminal.

Install the desktop build for macOS, Windows, or Linux, or contact us directly for institutional evaluations, partnerships, and platform enquiries.

Downloads Contact Us