AI Code Detector
Estimate whether source code was written by an AI assistant.
Paste at least 30 lines of code. A code-specialised open-source model measures how predictable each function is, and suspicious functions are highlighted. This is the least reliable detector: use it as a conversation starter, never as proof.
AI Code Detector
1 credit
The verdict you get
Never a bare percentage: each verdict comes with a short explanation and the limits of the method.
What we look at
Free plan · 10 credits a month · No credit card
Paste the code
At least 30 lines; Python, JavaScript, TypeScript, Java, PHP, C/C++ and more.
Function-by-function scoring
Each function gets its own score and style cues are listed.
A verdict in three levels
Probably human, uncertain or probably AI — never a bare percentage — with a short explanation and the limits of the method.
When to use the AI Code Detector
A take-home coding test looks unusually polished
See which functions look generated before the interview.
A programming school reviews assignments
Get an indication to discuss with the student — never a sanction on its own.
We name our methods because a verdict should be explainable.
In your result
A result is an indication, not proof. Detectors can be wrong, especially on compressed, edited or very short content. Only submit content you have the right to share. If you think a verdict is wrong, tell us from the result page.
Other detectors
Estimate whether source code was written by an AI assistant.
Free plan: 10 credits every month. No credit card required.