Frequently Asked Questions
Answers to the most common questions about Data Utility: converting files, joining, masking, API tables, data lineage and privacy. Free, in your browser.
How big can a file be?
It depends on your device. A few hundred MB usually works.
Which conversions does the Format Converter support?
Any direction between JSON, CSV, Parquet and Excel (.xlsx), for one file or a whole folder. Nested JSON is flattened into columns such as address.city. For Excel, the first sheet is read; old .xls files are not supported.
Why did my number or date change after converting?
Types are detected from the values. Turn off "Detect types" to keep everything as text, which also keeps leading zeros in IDs and phone numbers.
Will large numbers change?
No. Unlike many tools, long numbers like 12345678901234567890 are kept as they are.
What can the Data Formatter do?
Validate, beautify, minify and clean JSON, CSV, XML and YAML. Errors show the exact line and column, and JSON can be auto-fixed for common mistakes such as trailing commas and single quotes.
What is gzip?
A common compression format. Most tools, databases and programming languages can read .gz files.
How do I open a .gz file?
Use the Decompress tab here, or double-click it on Mac and Linux, or use 7-Zip on Windows.
What does the File Compressor do?
It shrinks files with gzip (.gz) in your browser and opens .gz files. Text formats such as CSV, JSON, XML and YAML often shrink by 70 to 90 percent. Parquet is already compressed, so it shrinks only a little.
Is my image uploaded?
No. Only the OCR engine is downloaded once; your image stays on your device.
Does it read handwriting?
No, only printed or on-screen text.
Can I convert many screenshots at once?
Yes. Choose Many screenshots, drop all the images, pick CSV, Excel, JSON or Parquet, and download one file per image in a ZIP.
How accurate is Screenshot to Data?
It uses OCR in your browser, so accuracy depends on the picture. Sharp, high-contrast screenshots work best. Always check numbers and decimal points against the image before you use the data. The image is read on your device and is not uploaded.
Does it connect to my database?
No. It only writes the SQL for you to run yourself.
Which databases does the SQL Generator support?
PostgreSQL, MySQL and MariaDB, SQL Server, Oracle and SQLite. It writes CREATE TABLE and INSERT statements from a CSV, JSON, Excel or Parquet file, and guesses column types from your data.
Which databases are supported?
CREATE TABLE from PostgreSQL, MySQL, SQL Server, Oracle and SQLite. Comments, quotes and schema names are handled.
Are foreign keys checked?
No. A foreign key column gets a number from 1 to 100, so load the parent table with ids 1 to 100 first.
Which JSON Schema versions work?
Draft 7 and newer for the common keywords listed above. Unknown keywords are ignored.
Is the data real?
No. It is random and fictional, so it is safe to share.
Can I get the same data again?
No – each click makes new data. Download to keep a set.
How many rows can the Data Generator create?
Up to 500,000 rows at a time. Choose column types such as names, emails, addresses and dates, then download as CSV, JSON, Parquet or Excel.
What counts as a duplicate row?
A row identical to an earlier one in every column.
What does the Data Profiler show?
Column types, missing values, duplicates, outliers and basic statistics, so you can judge the quality of a file at a glance. You can download the report, with an optional password.
Can the files be different formats?
Yes. A can be a CSV and B an Excel file. Columns are matched by name. JSON and Parquet files are not supported here, so convert them to CSV first.
How does Data Comparison find differences?
It matches rows by a key column, or by the first two columns, and shows which rows were added, removed or changed, down to the cell. It works with CSV and Excel files.
What is sent to Google Gemini?
Column names, types, statistics and, if you allow it, the first 5 rows and the most common values. Your full file is never sent.
Which AI does it use?
Google Gemini. You use your own Gemini API key, which you can get free from Google AI Studio.
Can I get a report or download the chat?
No. The chat gives answers only. It cannot create reports or export data, and nothing is saved.
What happens if I misuse the chat?
Off-topic, abusive or manipulative requests are blocked. After 3 violations the chat pauses for 5 minutes.
Are the column types correct?
Types are guessed from your data in your browser, not by the AI. Check them before you run the script.
What is the Data Dictionary Generator?
It writes a CREATE TABLE script with a description for every column, for PostgreSQL, MySQL, SQL Server, Oracle or SQLite. It uses your own Gemini key and sends only a summary of your data.
Will the README be accurate?
It is based on statistics from your data. Read it once before you share it, because the AI may guess the meaning of unclear columns.
What is the Data README generator?
It writes a README.md that explains what your dataset is about, with a table of columns, based on statistics from your file. Read it once before sharing, because the AI may guess the meaning of unclear columns.
What is data lineage?
It is the history of a file: where it came from and every step that changed it. Here you can see which tool did what, in order.
Where is the history stored?
Only in your browser, next to your Recent files. It is never uploaded, and deleting a file deletes its history.
What is Data Lineage?
Data Lineage shows the history of your files as a graph: which tool masked, converted or cleaned a file, and what it did. Click a step to see details, such as which column was masked and how. It is saved in your browser next to Recent files and is never uploaded.
Where is my Data Lineage stored?
Only in your browser, next to Recent files. It is never uploaded. You can name each lineage, and deleting a file deletes its history.
How does the token work?
Choose Bearer token, API key in a header, or username and password, and paste the value. It is sent only to the API you call, kept in memory, and never saved or written to Data Lineage.
What if the API blocks the request?
Some APIs do not allow calls from other websites (CORS), so the browser blocks the request. Use an API that allows browser access, or ask its owner to enable CORS.
What happens to nested fields?
Nested objects are flattened into columns such as address.city. If the response has several lists, you can choose which one becomes the table.
Is it safe to use a token in API Response to Table?
Your token goes only to the API you call, straight from your browser. Data Utility never receives it or the response. It is kept in memory for that request, not saved, and not written to Data Lineage. Use a token with read-only access when you can.
Can I merge different formats?
Yes. You can mix CSV, JSON, Excel and Parquet files. They are read, stacked one after another and saved in the format you choose.
How can I split a file?
Choose rows per file, a number of equal parts, or one file per value of a column. You download all parts as one ZIP, in CSV, JSON, Excel or Parquet.
What if the files have different columns?
All columns are kept. Cells for a column that a file does not have are left empty, and the tool lists those columns so you can check.
Can I merge CSV and Parquet files together?
Yes. You can mix CSV, JSON, Excel and Parquet files. They are stacked in the order you add them and saved in the format you choose. Columns missing in some files are left empty.
What kinds of text can I paste?
Tab, comma, semicolon, pipe or space separated text, Markdown tables, and key: value lines. The separator is detected for you and you can change it.
Is my text uploaded?
No. The text is turned into a table in your browser and is never sent anywhere.
What text can Text to Table read?
Tab, comma, semicolon, pipe and space separated text, Markdown tables, and report-style text. Blocks of Key: value lines separated by blank lines become one row each, and sentences such as "Order #1001 was placed on 2026-10-20 by customer Rahul. Product: Laptop. Status: Delivered." become rows with columns for the repeated parts. The layout is found automatically, and you can pick it yourself if the columns look wrong. If the text mixes several layouts, the tool tells you instead of making a table that could be wrong.
Why does Text to Table say the text mixes layouts?
It only builds a table when the lines or blocks look alike. A paste that contains a log line, a JSON line, a sentence and a Key: value block together has no single set of columns, and any table made from it would be a guess. Paste one layout at a time, or choose the layout yourself in the list.
What is synthetic data?
Data made by a program that follows the patterns of real data, such as ranges, shares and how columns move together, without containing the real records.
How is it different from the Data Generator?
The Data Generator makes random rows from column types you choose. The Synthetic Data Generator learns from your own file and makes rows that match its patterns, so the results look like your real data.
Is synthetic data fully anonymous?
It lowers the risk, and no real row is copied, but it is not a guarantee. Review it before you share it publicly.
How is the Synthetic Data Generator different from the Data Generator?
The Data Generator makes random rows from column types you choose, such as names, emails and dates, so the columns do not relate to each other. The Synthetic Data Generator learns from your own file and makes rows that match its ranges, shares and the way number columns move together, so the result looks like your real data. Use the Data Generator for quick fake data, and the Synthetic Data Generator when the data should behave like production.
Which transformations are available?
Join with a second file (inner, left, right, full), trim spaces, concatenate columns, change case, find and replace, split a column, rename, remove columns, filter rows, remove duplicates, sort and fill empty cells.
Can I run several steps?
Yes. Steps run from top to bottom and each one sees the result of the one before. You can reorder or remove them.
Can I join two files or combine columns?
Yes. Data Transformation lets you join two CSV or Excel files on a key (inner, left, right or full), and trim spaces, concatenate columns, change case, find and replace, split, rename, filter, sort, remove duplicates and fill empty cells. You choose the steps, check the result and download it. Your files stay in your browser.
Can I change the date format of a column?
Yes. Add a Change date format step, pick the column and the new format, such as 2026-03-09, 09/03/2026 or March 9, 2026. It reads many layouts, Excel date numbers and unix timestamps, and can also give only the year, month, weekday or quarter. If 03/04/2026 is unclear, you choose day first or month first.
Is masking the same as anonymizing?
Not always. Combined details can still identify people, so review before sharing.
Can a hash be reversed?
Not directly, but easy-to-guess values can be matched. Use a salt.
How does Data Masking protect personal data?
It finds columns that look like emails, phones, names or card numbers, and lets you choose a method for each: fake data, partial masking, hashing, noise, rounding or removal. The same input can give the same output, so relationships between files still line up.
Can I prove a file was masked?
Yes. Save the masked file to Recent files and open Data Lineage. It lists every masked column and the method used.