How to use it
- 1
Open an authenticated project and run a bounded crawl.
- 2
Wait for the stored result and open the duplicate groups.
- 3
Review the intent of each group manually: identical metadata is not always a defect.
What the tool can do
- Groups exact matches of non-empty titles and descriptions separately.
- Shows URLs from the stored result without making another network request.
- Uses the same ownership-protected result as the project crawl overview.
Common use cases
- Find template-level title duplication in the fetched category or service sample.
- Review repeated descriptions after a CMS template change.
- Prepare a URL list for manual editorial review.
How it works inside
Grouping uses exact stored text; semantic similarity and AI classification are not used.
Empty values are not duplicate groups and should be reviewed separately as missing metadata.
Questions and answers
Does it check every page on the site?
No. Groups are built only from one bounded crawl, with at most 25 HTML pages.
Is similar but non-identical text considered a duplicate?
No. The current implementation groups exact matches and does not present heuristics or AI judgment as fact.