An AI credit is the unit metering analyses generated by artificial intelligence. The important point fits in one sentence: no statistic in this glossary consumes any. Figures are computed from your trades by fixed formulas, and AI is only there to comment.
Contents
An AI credit counts one request sent to an artificial intelligence. Each request consumes a variable number depending on how much data has to be read: a review of three hundred trades costs more than a question about one specific setup.
The meter exists for a simple economic reason, and it is better said plainly. The rest of the journal has a zero marginal cost: recomputing your profit factor a hundred times costs no more than computing it once. An AI call does not work that way: it is billed every time, by volume.
The only other option would have been to quietly degrade the quality of answers as usage rises, which many tools do without saying so. A visible meter is less pleasant and more honest: you know what is left, and the answer you get is always the same one.
How you spend changes a lot about what you get. A global analysis of the whole history costs several credits and returns a general summary, often accurate and rarely actionable. A precise question about a setup, a market or a time window costs less and returns something you can apply the next day.
And the useful habit is to look at the statistics first, since they are free. The figure tells you what to look at, the AI helps understand why. Asking the AI to find the problem when the table already shows it amounts to paying for a reading you could do yourself.
The classic mistake. Asking for a global analysis when a precise question would do. A review of the whole history costs several credits and returns a summary that looks like what you already knew; a question about a setup, a market or an hour costs less and returns an answer you can apply. The free table says what to look at, and AI only comes after.
They meter analyses generated by artificial intelligence: a trade review, a period summary, a coach answer, a finely analysed check-in. Each request consumes a variable number depending on how much data has to be read.
No, none. Net P&L, win rate, profit factor, drawdown, the Oshi Score and everything else are computed from your trades by fixed formulas, with no AI and no cost. Your imports and your charts do not consume any either.
Because the rest has a zero marginal cost: recomputing a statistic a hundred times costs no more than computing it once. An AI call is billed every time and by volume. The other option would have been to quietly degrade answer quality as usage rises, and a visible meter is more honest.
AI analyses stop until renewal, and nothing else stops. Your journal, your statistics, your imports and your charts carry on exactly as before. Renewal lands on your subscription anniversary date, not on the 1st of the month.
By asking precise questions rather than requesting global analyses, and by looking at the statistics first, since they are free. The table tells you what to look at, and AI then helps understand why. The reverse order costs more and returns less.
The meter sits in Settings, under Subscription. It shows what has been used in the current period, the next renewal date, and any extra balance you bought, which does not expire and is used after the monthly quota.
Each plan has its quota, and the free tier has zero, which is a decision and not an oversight: AI analysis is, along with automatic syncing, one of only two items that cost per use. A free account therefore sees the preview screen rather than a handful of trials.
To go deeper on this: read the full guide. Every term is gathered in the glossary.