CleanMySheet guide
CleanMySheet groups similar rows using normalized Levenshtein similarity (default ~85%) after blocking on a short key prefix so large files stay usable. Matches appear in a review drawer; nothing is removed until you mark a pair as a duplicate.
Near-duplicates wait for your review.
Drop CSV or Excel here
Tap to choose a file. Parsing stays on this device.
Parsing stays on this device. Optional: drop a .cleanmysheet.json recipe with it.
Steps #
- Run a scan after upload
- Add Fuzzy duplicate detection from search or Recommended
- Open Review fuzzy pairs
- Mark Duplicate or Keep both
- Undo anytime
Not an LLM guess #
Similarity is a string distance, not a model hallucination. You can undo the whole step from the recipe list.
Related searches this page answers #
- fuzzy duplicate finder
- near duplicate rows
- similar contacts csv