A table you can understand.
This CSV viewer opens supported files locally so you can read rows and columns without installing spreadsheet software. Search, sort, filter and keep a useful copy. Your original file is never edited.
Open a CSV file
Choose Open file or Paste data. CSV normally uses commas; TSV uses tabs. File settings lets you choose a delimiter, character encoding and whether the first row contains column names. The header choice is visible and reversible.
JSON collections and supported macro-free XLSX sheets can also be viewed. Workbook formulas are not calculated and stored results can be out of date.
Understand the view
Search checks the visible columns. Filters narrow the rows, and column-header controls change their order. Source row numbers stay with each record. Hiding a column does not delete it; selection counts distinguish matching rows from selections hidden by filters.
Column summary and Review data describe values and structure. They do not certify business data, fix errors or validate a schema.
Keep the result
Export all matching rows, this page or selected matching rows. Choose visible columns or all columns. CSV and TSV are useful table copies; JSON keeps a positional view report, and Markdown makes a readable table or review summary.
Generated copies can change headers, fill missing cells or prefix formula-like values according to the settings you choose. An original-file download preserves the supplied bytes.
Why columns can look wrong
A comma inside a quoted field is part of that cell. A quoted newline is still part of the same record. If a table appears in one column, check its delimiter. If characters look wrong, choose the file’s encoding. Duplicate or empty headers get positional labels while their original text is preserved. Uneven rows are retained and reported.
See three examples of wrong-column problems or read the full Manual.
A small example
Name,City,Note
Ada,"New York, NY","Line one
Line two"
Lin,Paris,"He said ""hello"""
This contains two data records. Ada’s city includes a comma, her note includes a newline, and Lin’s note includes a quotation mark. Filter City for Paris to export Lin’s one matching row.
Files, limits and privacy
Delimited text is limited to 10 MiB; JSON/JSONL and XLSX to 5 MiB. A table can contain up to 100,000 logical rows, 200 columns and 1,000,000 normalized cells. Browser memory still matters. The Manual lists all bounds and supported formats.
Your table, filters and selections exist only in this page’s memory. Closing or leaving can lose the view. Copy or download what you need. This app does not upload files or save work automatically. Website hosting and the shared appearance preference have separate boundaries described in Privacy.
Can I edit cells or calculate workbook formulas?
This is a read-only viewer. It does not edit cells, write XLSX workbooks, run macros or recalculate formulas. Stored workbook values may differ from a freshly calculated spreadsheet.
Why do exports mention formula-like values?
Spreadsheet programs can interpret some text as formulas. Quoting a CSV field alone does not prevent that. The prefix option changes affected values to reduce risk; it is not a universal safety guarantee. Review the policy before opening exported data in another program.
Need to record an observation? Open Notepad and explicitly copy your notes. No table data transfers automatically.