Real-Time Lucky Jet Predictor

I want to build a small desktop or web tool that gives me real-time predictions for 1Win’s Lucky Jet game and its Mines variation. The core of the job is an algorithm that reads the live stream of crash multipliers, learns from historical rounds, and then outputs an on-screen indicator a few seconds before the next take-off.

To make the numbers trustworthy, the system has to:
• pull or accept past-round data automatically, run statistical analysis, and refresh its model continuously;
• display the current confidence level and suggested cash-out point in an intuitive interface (a simple panel or overlay is enough);
• include a parallel module for Mines that estimates safe tiles in real time using the same data ingestion pipeline.

Real-time prediction is the priority; daily or weekly summaries are optional only if they help the main goal. I’m open to the language or framework you prefer—Python with pandas, a lightweight Node/React stack, or anything that can sit beside the 1Win site without violating its terms. All I need back is:
1. the working predictor with its UI,
2. commented source code, and
3. short setup instructions so I can run it on my own machine.

If you’ve tackled crash or Mines predictors before, show me a quick demo or repo link and let’s get started.