AI CODE DETECTOR GUIDE
Does an AI code detector give the same score twice?
With the current Is My Code AI rules, yes: scanning the same source text again produces the same matched signals and total rule weight. The browser applies fixed text-pattern rules rather than drawing a fresh random result for each scan.
Analysis runs in your browser; submitted code is not uploaded.
What “the same input” means
The detector evaluates the source text against its configured rules. If that text and the rules stay the same, the same rules match and their weights add to the same score. Changing a comment, identifier, error message, or other source text can change which signals match.
The language picker does not supply a different scoring model to the current analyzer. The score is a sum of matched rule weights, not a probability that AI wrote the code.
How to compare two saved results
Run the same unchanged snippet twice, then compare the matched-signal list and score in each report. The downloadable JSON also includes the declared ruleset version; the current value is 1.0.0. Keep that field with the result when comparing scans taken at different times.
If the code or configured rules changed between scans, a different result may be expected. Check the exact input and the listed matches before comparing the totals.
A repeatable score is still only a heuristic
Repeatability means the same configured checks respond consistently to the same text. It does not validate the checks as a way to identify an author. A stable match cannot prove that a person or an AI tool wrote the code.
Open the AI code detector and review what its score represents or how to inspect matched lines.