Zain Dana HarperThe Quanta products

PRODUCTS · THE QUANTA SUITE

Standalone tools you can actually run today.

A suite of practical software, separate from the accountability research: a professional display-calibration app, a set of forecasting and paper-trading toolkits, and a shared interface layer that ties them together. Each one is a real tool with real tests, built to the same habit of saying plainly what it does and what it doesn’t. They don’t depend on the research — they stand on their own, and they help fund the slower work.

Honest up front: these are v1.0.0 beta, distributed as source (and, for calibrate-pro, a standalone Windows build) — not yet on a package index. Tested and usable; still maturing.

six products · v1.0.0 beta · ~830 tests across the suite · source & standalone builds

calibrate-pro — the most finished of them.

A professional display-calibration suite for Windows, with both a graphical interface and a command line. It detects your monitors, matches the panel against a database, and applies real corrections through the hardware — DDC/CI control (sending commands directly to the monitor over a standard cable) and 3D-LUT output (a color-correction table loaded into the display), ICC v4 profiles, HDR-aware. It has a sensorless mode that predicts a good profile from a database of 58 characterized panels (predicted ΔE under 1.0), and a native i1Display3 driver that talks to the colorimeter directly over USB, no third-party color stack required. This is the genuinely shippable one: 297 tests, ~26 working CLI subcommands, and a standalone .exe so you don’t need Python at all.

The color-science origin: years of film LUT pipelines and display work, packaged into a tool a person can actually run.

The forecasting & trading toolkits — paper-first, by design.

quanta-oracle — a toolkit for predicting trends in data over time and spotting unusual breaks or jumps: ARIMA with automatic order selection, structural-break (changepoint) detection by information criterion (a score that picks the simplest explanation that still fits the data), and feature engineering — building extra signals like recent averages, lag values, and Fourier terms to help the model see patterns. 187 tests. The honest forecasting layer the others build on.

quanta-finance — a rule-based trading toolkit for stocks and crypto: backtesting (running a strategy against old data to see what would have happened), portfolio optimization using named methods (Markowitz, Black–Litterman, Hierarchical Risk Parity, Risk Parity), and strategies like an EMA-crossover momentum filter. 142 tests. It runs the analysis and the backtest; it doesn’t pretend a backtest is a promise.

quanta-engine — the self-adjusting loop on top: it trains forecasters, turns forecasts into signals, paper-trades or backtests them, and feeds the results back to adjust its own internal settings (the numbers that determine how the model behaves). 178 tests. It is paper-trading by default, and live brokerage is deliberately hard to switch on — it refuses to run live unless you set an explicit risk-acknowledgement in the environment. The guardrail is the feature.

These are real toolkits, honestly beta. Nothing here is financial advice, and the design assumes you’ll paper-trade first — it makes that the path of least resistance on purpose.

The suite, and one tool reporting itself.

quanta-ui is the shared visual style and controls (a PyQt6 theme and widget set) the GUI apps are built from, so the suite looks like one thing. quanta-ecosystem is the single package that installs the whole set in one command, and quanta-color is the color-science member, with its own page. Here is the most mature member confirming what it is.

Exhibit I calibrate-pro · the CLI surface is real, not aspirational
$ python -m calibrate_pro --version
Calibrate Pro v1.0.0

$ python -m calibrate_pro --help
Calibrate Pro - Professional Display Calibration Suite
positional arguments:
  {detect, calibrate, verify, list-panels, list-targets, match, refine,
   ddc-calibrate, ddc-info, native-calibrate, hdr, uniformity,
   export-panel, import-panel, generate-profiles, gui, status, ...}
    detect            Detect connected displays
    native-calibrate  Calibrate using the native i1Display3 driver (no ArgyllCMS)
    ddc-info          Show DDC/CI capabilities for connected monitors
    verify            Verify calibration accuracy

The breadth is real: detection, hardware DDC/CI control, a native colorimeter driver, verification reports. Not a wrapper around someone else’s tool — the driver and the panel database are the work.

Real output from calibrate-pro v1.0.0 (297 tests). Suite test counts: calibrate-pro 297 · quanta-oracle 187 · quanta-engine 178 · quanta-finance 142 · quanta-ui 17 · quanta-ecosystem 10. Honest limit: every member is marked beta and is currently installed from source (calibrate-pro also ships a standalone Windows build) — they are packaged for a public index but not published to one yet, so I won’t claim “available on PyPI.” calibrate-pro is the one I’d call genuinely product-grade today; the quant tools are substantive but still beta.

Where they really are.

These are real, tested software you can use today — a product line that also funds the slower accountability and language work and gives it somewhere practical to land. calibrate-pro is the most finished — a tool I’d hand to a colorist today. quanta-oracle, quanta-finance, and quanta-engine are substantive beta toolkits with named, tested methods, built paper-first with their risk gates on by default. quanta-ui and quanta-ecosystem are exactly what they sound like — the shared look and the one-command installer — small glue, not headline products.

Tested and usable, honestly beta, and not on a package index yet — I’d rather you know that than find out.

GitHubHarperZ9  ·  Color memberquanta-color →  ·  ← the map  ·  the index