Every source that overlaps the document, with how much it covers on its own. In the match overview each word is counted once, against its largest source, so the shares there add up to the similarity index.
# Source Overlap Counted
1
Example journal article A
Example Journal of Agronomy · 2017
12.0% 23 words
12.0%
Side by side
· Exclude
2
Example journal article B
Example Journal of Soil Science · 2021
11.0% 21 words
11.0%
Side by side
· Exclude
3
Example conference paper C
Example Conference on Farm Economics · 2019
5.2% 10 words
5.2%
Side by side
· Exclude
Plagiarism 28.3% · 54 words that need attention
54 words matched with no citation nearby (passages of 8+ words)
0 words cited but copied word for word at length (40+ words, no quotation marks)
Not counted: 0 cited words, 0 in quotation marks, 0 in short uncited phrases (names, stock phrases)
AI-writing signals 32.6% · Many passages read as machine-generated (sample)
62 of 190 words are in passages that read as machine-generated. A signal for review, not proof.
See passages
UGC 2018: Level 1 (above 10% to 40%)
Revise the manuscript
What the levels mean
Level 0 up to 10% Level 1 above 10% to 40% Level 2 above 40% to 60% Level 3 above 60%
UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018. Your institution decides how they apply.
Similarity is the share of checked words that also appear, as runs of 5 words or more, in the listed sources. Each matched word is counted once, against its largest source. A similarity index is evidence for a reviewer, not a verdict of plagiarism.