CSV, JSON, Excel, YAML & XML Converter

This converter moves data between the formats people trade most: CSV, TSV, JSON, JSON Lines (NDJSON), YAML, TOML, XML and Excel workbooks. Paste or drop a file and it works out which format you have; pick another and the result appears straight away. It can also write SQL INSERT statements for PostgreSQL, MySQL, SQL Server, SQLite or standard SQL, so a spreadsheet becomes a database load script in one step.

CSV follows the RFC 4180 rules spreadsheets use, so quoted commas and line breaks inside a cell are read correctly. Nested JSON is flattened into dot-named columns such as address.city, and rebuilt when you convert back.

Open an .xlsx file and pick any sheet; save any result as a real Excel workbook with typed numbers and a bold header row. When something is wrong, you get a plain-English message with the line and column and a button that jumps to it. Everything happens in your browser, so private exports never leave your device.

1. Paste your data into “Your data”, press “Open file”, or drop a file on the box: .csv, .tsv, .json, .ndjson, .yaml, .toml, .xml or an Excel .xlsx (then pick its Sheet). “Example” loads a sample.
2. Leave “From” on Auto-detect (the badge shows what was found), or choose the format yourself.
3. Under “Convert to”, pick JSON, NDJSON, CSV, TSV, YAML, TOML, XML, SQL or Excel. For SQL, type the Table name and choose the Database.
4. Check the options: “First row is headers”, “Detect numbers, true/false and null”, “Trim spaces”, “Skip empty rows”, and “CSV settings” for delimiters, quotes and how lists are written.
5. Switch the result to “Table” to sort and filter the rows, then press “Copy” or “Download”. The round arrow button swaps the result into the input to convert it back.

• Seven text formats in any direction: CSV, TSV, JSON, NDJSON (JSON Lines), YAML, TOML and XML, plus format-to-same for tidying.
• Excel .xlsx in and out: open any sheet of a workbook (dates come in as 2026-10-07), and download results as a real workbook with numbers kept as numbers and a frozen header row.
• SQL INSERT output with the right quoting for each database, an optional CREATE TABLE with column types worked out from the data, and many rows per INSERT or one per row.
• Auto-detects the input format, and the CSV delimiter: comma, semicolon, tab or pipe.
• RFC 4180 CSV reading (quoted commas, line breaks in cells, Excel’s byte-order mark), with Trim spaces and Skip empty rows switches.
• Optional type detection: 42 becomes a number, true/false a yes-no value, empty cells null. Codes such as 007 stay as text.
• Nested JSON flattened to dot-named columns, with lists as columns, joined in one cell, or kept as JSON text; dot headers rebuilt into nesting on the way back.
• XML attributes read as @name and text beside them as #text; repeated tags become lists. TOML tables and [[arrays of tables]] read and written; dotted keys read.
• Table preview with sortable columns and a row filter, and plain-English errors with “Go to the mistake”.
• Copy, download with the right extension (.toml, .ndjson, .sql, .xlsx…), and an Excel-friendly CSV option.

• Excel in some countries uses semicolons, because the comma is their decimal mark. If your CSV opens in one column, choose “Writing: separate with Semicolon” and tick the Excel-friendly download.
• Turn off “Detect numbers, true/false and null” when every value must stay exactly as typed, such as product codes like 1E5.
• TOML has no null and must start with named keys, so empty values are left out and a plain list is put under “rows” — the status line says when this happens.
• For a big SQL load, keep “Many rows per INSERT” ticked: 500 rows per statement runs far faster.