Understand AI Grading, Hints, and Passing
Scripta does not score only character overlap with the source. It judges whether your answer performs the same meaning and role, then identifies what to change next.
The default pass threshold is 80
A cortex begins at 80 and you can change the threshold in settings. A score at or above it passes only when no hint was used.
Use meaning gaps, awkward phrasing, and strengths in the feedback rather than treating the score as the whole result.
Every level is graded semantically
Sentence grading looks at phrasing and meaning chunks; paragraphs add flow; whole texts add structure and missing central moves.
Scope and AI cost change with level. Check the current amount where you submit or on pricing.
A hint helps but does not pass
Using full meaning, structure, or a word hint prevents that attempt from passing even when its score is high.
Complete it with help, read the feedback, then return and reproduce it unaided.
Give up is a free exit
Give up reveals the source without calling AI grading, so it uses no credits. It also records no score or pass.
You can move on after a failed item; it remains available in a later queue.
Passed items shrink the queue
Scripta does not repeat every item equally. Passed work leaves the default queue so practice narrows toward what still stalls.
A session remembers the item where you left. If it has since passed and disappeared, the remaining queue begins at its first item.
A score is a signal for the next attempt, not the end of the practice.