How to open CSV and Excel files privately in your browser

CSV and TSV files are delimited text. They usually contain one table, have no universal schema, and depend on choices such as character encoding, delimiter, quoting, and whether the first record is a header. An XLS or XLSX workbook can contain many worksheets, formulas, cached results, formatting, merged cells, charts, comments, and other document structure.

Choose the viewer from the source format rather than from how the file looks after opening. A CSV viewer is the better diagnostic tool for separators, uneven rows, text encodings, and column profiling. An Excel viewer is the better choice when worksheet boundaries, cell coordinates, or stored formulas matter.

A browser can read a file you deliberately select through the File API and process its bytes in JavaScript. A local viewer can therefore parse a CSV or workbook without sending the selected file to the site’s application server. Closing or refreshing the page normally clears the in-memory parsed data unless the tool explicitly stores it.

That is a narrower claim than complete confidentiality. The page still came from a remote origin, and browser extensions, device malware, clipboard history, screen sharing, crash reports, analytics unrelated to file contents, and cloud-synced downloads may create other exposure paths. Use approved offline software for regulated or high-impact records.

Start by checking the detected delimiter and text encoding. Commas inside quoted fields should not create extra columns, doubled quote characters may represent a literal quote, and quoted fields can contain line breaks. A row with a different field count may indicate bad quoting, the wrong delimiter, or a legitimately irregular export.

Next confirm whether the first row contains headers, search representative values, sort carefully, and inspect inferred column types. Type inference is descriptive rather than authoritative: account IDs with leading zeros, localized dates, currencies, and large numbers can all be mistaken for another type if the surrounding schema is unknown.

Confirm the filename, extension, sheet names, and used ranges before interpreting the grid. Switch through every relevant sheet and compare displayed cell values with stored formulas where available. A formula result may be a cached value written by the last spreadsheet application that calculated the workbook; a lightweight viewer does not necessarily recalculate it.

A data-grid viewer is not a visual clone of Excel. Charts, comments, conditional formatting, precise dimensions, hidden content, merged layouts, external links, macros, and accessibility structure may be omitted. Open important workbooks in a full compatible editor when those features affect meaning.

CSV export flattens a table into text. It cannot preserve multiple worksheets, formulas as executable workbook logic, formatting, charts, comments, validation, or other workbook features. A filtered CSV export may also contain only the analyzed or matching rows rather than the entire source.

Treat the export as a new derivative file. Reopen it, confirm the row and column counts, inspect quoting and non-ASCII characters, and verify identifiers with leading zeros. Keep the original workbook or CSV as the source of record unless your workflow explicitly replaces it.

Use a synthetic sample first, especially when testing an unfamiliar viewer. For real data, minimize the fields you expose, close unrelated extensions or screen-sharing tools, and avoid copying sensitive cells unnecessarily. Clear the page when finished and review where exported files are saved or synchronized.

For important decisions, compare a representative set of cells with the source application and document any row, column, file-size, or formula limitations. Local inspection is useful for quick review, but privacy and correctness both depend on the complete workflow around the browser.