Source
Verdict
Paste code on the left and press Analyze. Each matched heuristic lists the line numbers and snippets behind it; hover a finding to spotlight those lines in the editor.
Annotated samples
Most samples are curated — distilled from common LLM idioms so each one cleanly demonstrates one or two signals. A few are verbatim transcripts captured from popular models, to keep a real-world feel. Each card's badge tells you which. Load any sample into the analyzer to see how it scores.
Heuristics & weights
Each signal contributes its weight when matched. The score is the sum of matched weights over the sum of absolute weights, clamped to 0–100. Positive weights are AI tells; the single negative weight — Human-style complexity — is a counterweight that pulls the score down when a snippet shows genuine human variety.