Drop your file
Drop your CSV here
or browse files
drop another file to replace it
So we apply the right fixes and leave everything else alone.
Add a file and pick a profile to start
Need the file again?
Emails, countries, dates, prices, duplicates: fixed in one click.
Now do it with your file
Drop your file
Drop your CSV here
or browse files
drop another file to replace it
So we apply the right fixes and leave everything else alone.
Add a file and pick a profile to start
Need the file again?
Free plan cleans up to 1,000 rows. No signup. We delete your file about 10 minutes after cleaning.
What it does
Client lists with duplicate emails. Shopify exports with broken SKUs. Prices in three different formats. DataScrub fixes the cleanup work you keep putting off, so you send clean files instead of corrections.
How it works
Drag a CSV in, or browse for it. We clean it in memory and never store it.
Each profile already knows what that file type needs: emails, dates, prices, duplicates.
Fixed and formatted in seconds. Your download starts automatically.
Cleaning profiles
No prompts to write, no formulas to debug. Pick the profile and get the same result every time.
Fixes email typos and (at) syntax, normalizes countries and dates (flags ambiguous ones), removes duplicate emails and flags near-duplicates.
Normalizes prices across EU/US formats, cleans units, drops empty rows.
Flags SKUs Excel has damaged, converts dates to ISO and flags ambiguous ones, splits full names into first and last.
Drops aggregate columns, title-cases campaign names, repairs broken characters.
Normalizes phone numbers, fixes emails, removes duplicate contacts and flags near-duplicates. Numbers without an international prefix (+1, +44…) or that aren't valid are left unchanged and flagged in the phone_status column.
Your data stays yours
No signup, no login, no tracking. Open the page and clean.
We delete the cleaned file about 10 minutes after cleaning. Your upload is processed in memory and is not kept.
Deterministic rules only. Same input, same output, every time. Your data never trains a model.