Platform overview

Every stage of the loop, built as a real product surface.

Twenty-one pages in the workspace, five areas, one project carrying all of them. The six capabilities below are how the site explains it; each links to a deep page with real screens, the method as implemented, and a plain statement of what it does not do.

Get early access Why Quant Compose

5 templates

Research pipeline

Projects as isolated environments. Workflows composed from the catalogue, validated before saving, versioned on every save, re-run on new data with a per-step record.

Open

Workflow Composer pipeline view showing the connected research stages from market data to backtest.

119 methods

Method and model library

119 research methods in 19 categories, 72 checked against a reference. 43 machine learning models on one training path.

Open

QuantLab method library page: counts of 119 methods and 72 validated, a category list and method cards marked runs-here with validated or unverified badges.

4 split modes

Validation and evidence

Chronological split, train-only scaler, purge floored at the horizon, walk-forward folds with the worst fold reported, trials counted against you.

Open

Walk-forward validation panel: fold and embargo inputs, headline cards, a score-by-fold bar chart against the 0.5 coin-flip line and an interpretation column.

6 stress episodes

Portfolio and risk

7 construction methods, 4 VaR methods, 6 historical stress episodes replayed on your own data. Research and analysis, not execution.

Open

Exported efficient frontier figure: a frontier curve, a cloud of random portfolios coloured by Sharpe ratio, and markers for max Sharpe, minimum volatility and equal weight.

15 index catalogues

Bring your own data

A CSV contract, Yahoo Finance, FRED, Ken French, Alpha Vantage, Polygon and custom REST behind an allowlist. Preparation in a fixed, audited order.

Open

Full Data Preparation page showing cleaning and shape settings, all four leakage safeguards enforced, 18,478 rows reduced to 18,440, and the stages-applied table.

4 export formats

Reports and saved results

One record of what ran, exported as PDF, DOCX, HTML or JSON. A workbench that marks a block stale when its source has moved on.

Open

Second page of a Quant Compose PDF evidence pack showing the model section, out-of-sample validation figures including the worst fold, and the strategy backtest comparison.

Now / next / exploring / future

Roadmap

Implemented, planned, exploring, future. Only the first list is a claim about the software.

Open

The workspace, page by page

Five areas, twenty-one pages, in the order the sidebar presents them.

The names below are the on-screen titles. The eight-stage journey on the home page maps onto these areas; the platform pages above group them by capability.

Home

Workspace Dashboard
Projects & Workspaces
Workflow Composer

Market Research

Market Explorer
Data Preparation (Cleaning & Preprocessing)
Risk & Return
Portfolio Optimisation
Scenario & Stress Testing
Statistics & Exploratory Data Analysis (EDA)

Model Intelligence

Feature Engineering
Machine Learning (ML) Models
Model Forecasting
Tune & Experiment
Explainability

Strategies

Strategy Builder
Strategy Backtesting
Strategy Monitor

Quant Reports

Reports
QuantLab (Method Library)
Audit & Governance

Header

Settings & Connections
Search and command palette
Saved Results and Project Notepad

Strategy Monitor tracks strategies you have built and backtested. It is a record over the strategy specification and the backtest engine, not a connection to a broker. Private beta

Read the method first

Related capabilities

From QuantLab

Start with the sample project, then bring your own data.

Private beta is free. Access opens by cohort and every applicant hears back from a person.

Get early access Institutions and teams