TornX Torn stock market predictor
Source release / free forever
The final TornX release

The market engine.Now yours.

I’m quitting Torn, but TornX should not disappear with me. The complete source is now being released for free so the community can inspect it, run it, improve it, and host its own instance.

One repository. Cacher / research lab / predictor / authenticated dashboard / build scripts.

35Every listed stockOne complete market surface
15MChanged-state researchBuilt around market snapshots
∞Self-hostedYour machine / your data / your keys
What you’re downloading

More than a dashboard.

TornX is a complete stock-market research pipeline: it collects Torn market snapshots, studies distinct price changes, produces the next changed-state forecast, and presents the result through a private self-hosted interface.

01 / COLLECTCACHE

Historical market cacher

Collects and maintains timestamped market snapshots on the 15-minute grid, including a lightweight live append-only cache.

02 / RESEARCHLAB

Prediction laboratory

Tests changed-state behavior across historical windows and evaluates forecasting approaches without hiding the underlying measurements.

03 / FORECASTMODEL

All-stock predictions

Builds one forecast object for all 35 stocks, including expected prices, action signals, timing, and validation metadata.

04 / OPERATEWEB

Private web interface

Serves the TornX dashboard locally with authentication, live holdings, market-history sparklines, and fee-aware portfolio details.

The system

From snapshots to signals.

Every part is included so you can follow the data through the project, audit the assumptions, and change the implementation for your own use.

Input / 01Market cache

Timestamped stock prices form the historical and live data layer.

Engine / 02Prediction model

The research and predictor tools analyze the next distinct market state.

Output / 03TornX dashboard

Forecasts, holdings, signals, timing, and history meet in one interface.

Start with the repository

Read the README.
Run it your way.

The GitHub README contains the current requirements, file layout, configuration, build instructions, and setup flow. Read it completely before running the cacher, predictor, or public-facing server.

  1. 01Download or clone the complete repository.
  2. 02Review the README, requirements, and configuration.
  3. 03Prepare your local data and Torn API access.
  4. 04Run from Python or use the included build scripts.
No paywall / no subscription / no gate

Take TornX. Keep it alive.

The commercial chapter is over. The code is free. Fork it, study it, improve it, or host it for yourself; just read the documentation and understand what you are running.