CSV delimiter converter and repair
Fix semicolon/comma/tab files that import incorrectly.
Your inputs
Your results
Ready when you areYour results will appear here.
How it works
Convert comma, semicolon, tab, or pipe delimited data into a consistently quoted download. The parser handles doubled quotes and quoted multiline fields using the CSV quoting conventions documented in RFC 4180. Repair here means normalizing quoting and record separators; malformed quotes produce an error, and inconsistent column counts are reported without adding or discarding fields. Reference: https://www.rfc-editor.org/rfc/rfc4180. Related workflow: https://www.browserling.com/tools/csv-to-tsv.
Read the source reference ↗A few things to know
Automatic delimiter detection is a heuristic; review the preview or choose a delimiter explicitly. No header is inferred: the first record is included in counts and preview. Blank records are preserved. Input is limited to 1,000,000 characters, 10,000 records, and 200 columns per record. Quoted field line endings are preserved. This converter does not recover broken quotes, fix encodings, infer data types, or guarantee spreadsheet formula safety. All processing stays in your browser.
When to use this tool
Delimited files can fail to import because the receiving application expects commas while the source uses semicolons, tabs or pipes. This converter parses the cells and writes a consistently quoted download with your chosen separator. It also reports inconsistent record widths. Repair means normalizing valid structure, rather than guessing missing values or discarding malformed text. Inspect the preview before using the result in another application.
- Convert a semicolon export into comma-separated data for another system.
- Create TSV from CSV while retaining commas inside text fields.
- Review inconsistent row widths before importing a supplier file.
How to use csv delimiter converter and repair
Paste the original delimited text
Keep its quoting intact. Quoted cells may contain separators or line breaks. The first record is treated as data, not an inferred header, and is included in the record count.
Choose conversion settings
Select automatic detection or an explicit input delimiter, then choose comma, semicolon, tab or pipe output. Decide whether you need a UTF-8 BOM and whether apostrophe-prefixing formula-like cells is acceptable.
Inspect and download
Check the detected delimiter, column count and inconsistent-record warnings. The preview shows up to 25 records. Download the full converted file and import it using the matching separator in your destination application.
Convert a semicolon product list
Inputs
Use the example containing Product;Description;Price, a Notebook with description "Blue, dotted", and a Pen with a quoted hello description. Select automatic input detection and comma output, without BOM or formula prefixes.
Result
The converter detects semicolons, reads three records with three columns and reports zero inconsistent records. The notebook description remains one cell; the pen’s internal quotation marks are preserved with doubled quotes in the CSV.
What it means
Three records include the heading row. Quoting each output cell keeps the comma in the notebook description from becoming an extra column.
Common mistakes to avoid
Replacing separators with Find and Replace
A comma inside a quoted description belongs to the value. Structural parsing avoids replacing that content along with the actual separators.
Assuming mismatches were repaired
The tool preserves each record’s field count. A warning identifies rows to review; it does not invent missing cells or remove extra ones.
Ignoring changed formula-like values
Enabling prefixes adds a literal apostrophe, including before negative numbers. That can help spreadsheet handling but changes the exported cell text.
Frequently asked questions
Definitions, edge cases, and practical next steps for this tool. Search all tool FAQs ↗
Which delimiters can I convert?
Comma, semicolon, tab and pipe are supported for both input and output. Automatic detection tries those options and evaluates the resulting structure. If the result is ambiguous or the preview is wrong, select the known input delimiter explicitly.
What happens to commas inside quoted cells?
They stay inside the same field when the quoting is valid. The output quotes every field and doubles internal quotation marks. A description such as Blue, dotted therefore remains a single value in a comma-separated file.
Can the tool recover an unclosed quotation mark?
No. Broken quoting can make record boundaries uncertain, so the converter reports an error instead of guessing. Select the explicit input delimiter to see a precise parsing error, then correct the source quoting and rerun the conversion.
Does it remove blank records or trim spaces?
No. Blank records, empty fields and surrounding whitespace are preserved. The converter changes record separators and output quoting, not the meaning or cleanliness of the data. Use a separate cleaning workflow when those content changes are intentional.
What does an inconsistent-record warning mean?
A record contains a different number of fields from the first record. This may indicate a missing value, an unquoted separator or an intentionally variable structure. The download preserves that width; the preview marks missing fields rather than inserting them into the export.
When should I add a UTF-8 BOM?
A BOM can help some spreadsheet applications recognize UTF-8 characters. Choose it when your receiving application expects that signal. It does not repair already corrupted characters or convert an unknown original encoding; review accented and non-Latin text after importing.
Does formula protection make every spreadsheet import safe?
No. Prefixing suspicious values is a precaution whose behavior depends on the spreadsheet and import settings. It also affects negative numbers. Review the source, enable the option only when changed text is acceptable, and avoid opening untrusted exports without considering their content.
Are multiline cells and large exports supported?
Valid quoted multiline cells are supported and their internal line endings are retained. Input is limited to one million characters, 10,000 records and 200 columns per record. The preview is intentionally smaller than the download, so use the file for the full result.
Explore the topic in the blog
Sources and method
Use the references below to check definitions and platform guidance. The model’s scope and limits are explained above. Read our editorial standards.