The idea
One credit = one AI operation
Base statistics (delays, frequencies, draws) are included in every plan and consume nothing. Credits are used only when you ask the AI engine to compute something for you: a prediction with confidence score and explanation, a dream interpretation, an optimized system, a backtest.
The cost isn't arbitrary: it reflects how heavy the operation is. A prediction draws on already-trained models and costs 1 credit. Reading a photo (vision) or optimizing a system (MILP) requires far more compute, so it costs more. Same principle as a meter: you pay for the energy you actually use.
The price list
What each operation costs
Transparent price list, identical across all plans. Items with a dot are "premium" operations, more expensive to compute.
Run out of monthly credits? You can refill with a one-off top-up (30 / 100 / 500 credits) without changing plan.
The budget
It refreshes every week
Each plan includes a credit budget that refreshes EVERY WEEK (not at month end): you get an amount each week and unused credits don't roll over. It keeps usage steady instead of burning everything on day one. Below: the monthly reference and the weekly amount you refill.
You move up not just for more credits, but mainly to unlock tools: AI Advisor and backtests on Premium, expert tools (walk-forward, advanced systems, API) on Pro. Business/Enterprise plans work with NO credit limits: built for shops and professionals operating continuously.
What credits are NOT
Credits measure usage of the AI engine, not the probability of winning. No operation, however many credits it costs, raises draw odds or makes the game profitable: the expected value of lottery games stays structurally negative. Credits let you use analysis tools consciously.