TSV to JSON Converter

This tool takes tab-separated rows — the format you get when you copy a range out of Excel, Google Sheets, or Numbers, or export from a database client — and turns them into a JSON array of objects. Each data row becomes one object, with the first row supplying the keys. That makes it a practical bridge for the everyday gap between a spreadsheet and code: seeding a test fixture, preparing a small data file for an API request body, or handing a developer a structured version of a table someone maintains in a sheet.

It fits naturally into web and dev work, data prep, and office-to-engineering handoff. A common path is pasting a selection straight from a spreadsheet — clipboard data between cells is already tab-delimited — and getting back JSON you can drop into a config, a mock, or an import script. Students and admins use it the same way to reshape a small dataset without writing a parser.

Because everything runs in your browser and nothing is uploaded or kept after the tab closes, it suits a quick reshape of internal or sensitive tables — a staff list, a price sheet, an export you would rather not post to a server you do not control.

You provide TSV three ways: drop a .tsv (or tab-delimited text) file, paste rows directly, or load a sample to see the shape. The tool reads the first line as the header row, splits every line on tab characters, and builds one object per row keyed by those headers. The result is pretty-printed JSON — an array of objects, indented two spaces — which you then copy or download.

The parser is quote-aware: fields wrapped in double quotes are kept intact, doubled quotes inside a field are unescaped, and both Windows (CRLF) and Unix line endings are handled. If a header cell is blank it is named column_1, column_2, and so on by position, and if a row has fewer cells than the header it fills the missing ones with empty strings.

Two honest limits are worth stating up front. Every value comes out as a JSON string — the tool does not infer types, so 42, true, and an empty cell become "42", "true", and "", never numbers, booleans, or null. And it produces a flat array of flat objects; it does not build nested structures, group rows, or read multiple tables. It is a faithful row-to-record transform, not a schema designer.

The most frequent surprise is the header row. The first line always becomes the keys, so if your data has no header, the real first record is silently consumed as field names. Equally, duplicate header names collide — two columns both called name will overwrite each other in the object, leaving only the last value.

The second is the delimiter itself. Tabs are invisible, so data that was actually separated by spaces or commas parses as a single column per row, and a stray tab inside a value splits it unexpectedly. Confirm the source really is tab-separated before converting. Note too that everything stays a string: downstream code expecting a number or boolean must cast it, and a blank cell becomes "" rather than null.

Finally, watch row alignment and scale. Ragged rows pad short rows with empty strings and ignore extra cells beyond the header count, which can quietly misalign data if a tab is missing mid-row. Very large pastes are processed in the page and held in memory, so a multi-megabyte table can feel slow or strain the browser tab.

Open the JSON and confirm the object count matches your data rows (headers excluded), then read the first and last record end to end. Check that the keys are the column names you expected — not a real row promoted by mistake, and not blank column_N placeholders — and that no header was duplicated and overwritten. Spot-check a row with empty or quoted cells to confirm fields landed in the right keys and nothing shifted.

Remember the type rule when wiring this into code: numbers, booleans, dates, and blanks all arrive as strings, so validate or convert them where your application needs real types. If a column looks truncated or merged, re-check the source for an inconsistent tab or a value that contained a tab.

For anything official, financial, or legal — invoices, payroll, ledgers, records of regulatory weight — treat this conversion as a draft and verify the output against the source by hand, value by value for the fields that matter. The tool does not validate meaning, only structure.