AI Data Cleaner
Remove duplicates, fix inconsistent categories, normalise numbers and dates, fill missing values — all in one pass. Powered by DeepSeek AI.
Data Input
Cleaning Options
Remove Duplicates
Delete exact duplicate rows
Trim Whitespace
Remove leading/trailing spaces from all cells
Normalise Numbers
Strip $, £, commas → plain numerics
Standardise Dates
Convert all dates to YYYY-MM-DD
Unify Categories
AI: "nyc" + "NYC" + "New York" → "New York"
Rename Columns
Convert headers to snake_case
Remove Outliers
Drop rows >3 std dev from the mean
Fill Missing Values
Strategy for empty/null cells
Cleaned data will appear here
Configure options on the left, then click Clean Data
Your data is private
Data is sent securely to DeepSeek and never stored on our servers. We do not use your data to train AI models.
Frequently Asked Questions
What does 'Unify Categories' actually do?▼
This is the AI-powered operation. DeepSeek analyses all unique values in each categorical column and identifies near-duplicates: abbreviations (LA = Los Angeles), case variants (electronics = Electronics), typos, and shorthand. It canonicalises them to the most common form. Rule-based tools can't do this — it requires language understanding.
How large can my dataset be?▼
The AI cleaning path supports datasets up to ~12,000 characters, roughly 100–200 rows depending on column count. For larger files, clean a representative sample to validate your settings, then apply the same transformations to the full file.
What number formats are normalised?▼
Currency symbols ($, £, €, ¥), thousands separators (commas), percentage signs (%), and mixed formats. Example: '$1,234.50' → '1234.50', '45%' → '0.45'. The result is always a plain decimal string parseable by any data tool.
What date formats are standardised?▼
Common formats including MM/DD/YYYY, DD/MM/YYYY, Month DD YYYY (e.g. Jan 15 2024), and ISO variants are all unified to YYYY-MM-DD. The AI detects the format from context — e.g. 01/05/2024 in a US dataset becomes 2024-01-05.
What does 'Remove Outliers' do?▼
For each numeric column, rows where the value is more than 3 standard deviations from the column mean are removed. This is a statistical outlier definition (roughly 0.3% of a normal distribution). Use with care — enable it only if you know outliers represent data errors rather than genuine extreme values.
Is my data stored?▼
No. Your data is sent to DeepSeek's API for cleaning and immediately discarded. We never store your data on our servers and it is not used to train AI models.