AI-powered · Clean & Transform

Clean it.
Transform it.
Ship it.

Upload your Excel or CSV file, then describe what you need in plain English — fix inconsistencies, calculate new columns, split fields, reshape your data. Get a clean, transformed file back in seconds.

See how it works

// Free plan · No credit card needed · Clean + transform in one step

orders_q1_2024.xlsx
customer_name purchase_date amount product
john doe 3/5/24 $1,200.00 Office Chair - Exec
JANE SMITH March 5th 2024 850 exec office chair
Bob johnson 05-03-2024 $1,200.00 Office Chair
alice wong 2024/03/06 $975 Exec. Chair
Customer Name Purchase Date Amount Product
John Doe 2024-03-05 1200.00 Executive Office Chair
Jane Smith 2024-03-05 850.00 Executive Office Chair
Bob Johnson 2024-03-05 1200.00 Executive Office Chair
Alice Wong 2024-03-06 975.00 Executive Office Chair
Item Qty Unit Price Total VAT (12%) Grand Total
Executive Chair 4 240.00 960.00 115.20 1,075.20
Standing Desk 2 650.00 1,300.00 156.00 1,456.00
Monitor Arm 6 85.00 510.00 61.20 571.20

// 3 columns added · calculated from Qty × Unit Price, 12% VAT applied

Clean
Fix inconsistencies & errors
Transform
Reshape, compute & split
0 code
Plain English instructions

How it works

Upload. Describe. Download.

The whole workflow fits in three steps — upload your file, tell DataClean what you need done to it, and get a properly formatted file back.

01

Upload your file

Drop in any Excel (.xlsx) or CSV file. DataClean reads the structure, detects your sheets, and shows you a preview before anything happens.

.xlsx · .csv · up to 100MB on Max
02

Describe what you need

Tell DataClean what to fix or build in plain English — inconsistent dates, bad casing, duplicate rows, or something like "add a Total column from Qty × Price with 12% VAT."

clean · compute · reshape · split
03

Download clean data

Get a properly formatted Excel file back — clean headers, consistent values, light gray formatting. If something's off, give feedback and run another iteration.

Excel-in · Excel-out · preserved layout

Features

Built for cleaning and transformation.

Some data just needs cleaning. Some of it needs to be restructured entirely — new columns, split fields, calculated values. DataClean handles both in the same session.

Sheet selection

When your workbook has multiple sheets, DataClean lets you pick exactly which one to work on — it reads the full structure upfront and gives you control over what gets processed.

Iterative feedback

Not quite right? Tell the AI what to fix and run another pass. Each iteration builds on the previous one — whether you're refining a clean or adjusting a transformation.

Excel-in, Excel-out

Your file comes back as a properly formatted .xlsx — clean headers, auto-fit columns, and the same sheet structure. Whether you cleaned it, transformed it, or both.

Calculated columns

Add derived columns from existing data — totals, margins, tax amounts, percentage changes, running sums. Just describe the formula in plain English and DataClean builds it.

Split & merge fields

Split "Full Name" into First and Last, break an address into street, city, and ZIP, or merge multiple columns into one. No CONCATENATE formulas required.

Secure by default

Files are processed in isolated sessions and deleted after 24 hours on the free plan. Your data doesn't stick around longer than it needs to.

Real-world example

Not possible with Excel formulas

The kind of logic that usually needs a Python script.

This is a real transformation from an actual product catalog. Each kit (parent SKU) needs its total weight calculated by summing the weights of its component parts — but the parent rows don't know their children ahead of time, and the relationship lives across rows.

The instruction

dataclean · prompt
"In the Weight column, parent SKU rows have blank values. A row is a parent SKU when its SKU column equals its Parent SKU column.

For each parent SKU row, fill in the Weight by summing the Weight values of all rows that share the same Parent SKU value but are NOT the parent themselves (i.e. where SKU ≠ Parent SKU).

Leave child rows unchanged."

The result

SKUParent SKUDescriptionWeight
NEX3HT01parent NEX3HT01 Nexus 3PC Home Theater [KIT] —
NEX3HT011NEX3HT01Nexus 65" OLED TV18.5
NEX3HT012NEX3HT01Nexus Soundbar4.2
NEX3HT013NEX3HT01Nexus Subwoofer12.8
· · · 1,204 rows across 47 parent SKUs · · ·
VOR2GB01parent VOR2GB01 Vortex 2PC Gaming Bundle [KIT] —
VOR2GB011VOR2GB01Vortex Gaming Console3.9
VOR2GB012VOR2GB01Vortex Controller0.28
// parent Weight is blank · children have values
In Excel, this requires a nested SUMIFS with a self-referential exclusion condition applied only to parent rows — the kind of formula that takes serious Excel knowledge to write and even longer to debug. Most people who hit this problem reach for Python. DataClean solves it from a plain English description.

Pricing

Start free. Scale when you need to.

Three tiers based on file size and session count. Start on the free plan and upgrade when your usage outgrows it.

Free

$0

For individuals testing the waters. No credit card required.

  • 15 MB max upload size
  • 2 cleaning sessions per day (UTC)
  • Up to 10 iterations per session
  • Files deleted after 24 hours
Most popular

Pro

$ 25 /mo

For analysts and teams cleaning data on a regular cadence.

  • 50 MB max upload size
  • 100 sessions per month
  • Up to 15 iterations per session
  • Priority processing

Max

$ 85 /mo

For power users dealing with large, complex datasets every week.

  • 100 MB max upload size
  • Unlimited sessions
  • Unlimited iterations per session
  • Everything in Pro

FAQ

Common questions.

DataClean handles both cleaning and transformation. On the cleaning side: inconsistent date formats, mixed number formats, erratic capitalization, whitespace problems, misnamed columns, and duplicate rows. On the transformation side: calculated columns (e.g., Qty × Price + VAT), column splitting (e.g., Full Name → First + Last), column merging, and value standardization. If you can describe it in plain English, DataClean can likely do it.
Your files are only stored for the duration of your active session. On the free plan, sessions and their files are automatically deleted after 24 hours. On paid plans, session data is tied to your account and subject to the same deletion policy. DataClean does not share your data with third parties.
A session starts when you upload a file. An iteration is one round of AI cleaning within that session. You can describe different problems, review the output, and run another iteration until the data looks exactly right. On the free plan, you get up to 10 iterations per session. Pro includes up to 15 iterations per session; Max has unlimited iterations.
Yes. DataClean supports multi-sheet Excel workbooks and lets you choose which sheet to process. File size limits depend on your plan — 15 MB on free, 50 MB on Pro, and 100 MB on Max. For very large files, the Max plan is the right fit.

Clean and transform your next dataset in 60 seconds.

Describe what you want done to your data and DataClean handles the rest — cleaning, computing, restructuring.

View pricing