Logo

Ghost Scrub

Upload a dataset to begin

Supported formats: .csv,.xlsx,.xls,.xlsm,.xlsb,.tsv

Not signed in: Up to 20kB per day, no sign-up needed.
Sign in for free for unlimited files, no payment required.

About Ghost Scrub

Removes zero-width spaces and non-breaking spaces that break code.

Ghost Scrub scans your file for characters that render as blank but aren't, zero-width spaces (U+200B), non-breaking spaces (U+00A0), and other invisible Unicode marks that often sneak in from copy-pasted web content or exports from certain CMS and PDF tools.

These characters look identical to a normal space in a spreadsheet but break exact-match filters, VLOOKUPs, regex, and API calls downstream. The tool strips them out while leaving your visible text untouched.

Category: Cleaning & Preparation

Common Use Cases

  • Fixing lookup failures where two values look identical but a VLOOKUP or JOIN won't match them
  • Cleaning text copy-pasted from web pages or PDFs before importing into a database
  • Sanitizing form-submitted data before it's used as a matching key in another system

Key Features

  • Detects zero-width spaces, non-breaking spaces, and other invisible Unicode characters
  • Strips BOM markers left over from certain export tools
  • Leaves visible text and formatting untouched
  • Works across every column and cell in the file

Pair Ghost Scrub with Super Fix when a file has both encoding damage and invisible characters, or with Trim & Sanitize afterward to also clear empty rows and normal whitespace.

Frequently Asked Questions

What are "ghost" characters?

Invisible characters like zero-width spaces and non-breaking spaces that look like normal whitespace but break string comparisons, regex matches, and code that parses your CSV.

How would I know my file has this problem?

A classic symptom is a value that looks identical to another in a spreadsheet but fails an exact-match filter or lookup, that's usually a hidden character.

Essential for developers and data engineers debugging why an exact-match key comparison silently fails.

Read the related guide