100% client-side — files never leave your device

CSV cleaner

Clean spreadsheet exports before importing them: trim stray spaces, remove empty or duplicate rows, normalize column counts, and switch between CSV and TSV without uploading the data.

Drop a CSV or TSV fileOr paste rows below. Quoted commas and multiline fields are supported.

      How it works

      1. Open or paste the table. Choose CSV or TSV input; quoted separators and multiline cells are handled correctly.
      2. Select cleanup rules. Trim cells, remove blank rows, deduplicate exact rows and pad short rows independently.
      3. Review and export. The summary states exactly what changed before you copy or download the cleaned file.

      Clean the structure, not the meaning

      This tool deliberately avoids guessing whether “0012” is an ID, whether a date is day-first, or whether two differently cased names are the same person. Those are business decisions. It fixes mechanical issues that commonly break imports while leaving cell values recognizable.

      Check column alignment before importing

      A short row often means a missing trailing value; a long row can mean an unquoted comma. Normalization safely pads short rows but never deletes cells. If the reported column count surprises you, inspect the preview first. See the CSV cleanup guide for a reliable import checklist.

      Frequently asked questions

      Does it understand quoted CSV fields?

      Yes. Commas, escaped quotes and line breaks inside quoted fields are parsed according to RFC 4180.

      What does normalize column count do?

      It finds the widest row and pads shorter rows with empty cells. It does not discard extra values from wider rows.

      Is duplicate removal safe for a header row?

      Yes. Exact row deduplication keeps the first occurrence. A normal header differs from data rows and remains first.

      In what order do the cleanup rules run?

      Trim, then remove empty rows, then remove duplicate rows, then pad short rows. Because trimming happens first, rows that differ only by stray spaces count as duplicates; the summary states how many rows and cells each step touched.

      Does it catch near-duplicates or fix broken characters?

      No. Deduplication compares whole rows exactly and is case-sensitive, so “Ada” and “ada” both survive. A UTF-8 byte-order mark is dropped from the first cell, but the tool does not re-encode text or repair mojibake from an export that was saved in the wrong encoding — re-export those files as UTF-8.

      Related tools