Flatten Nested JSON

Nested JSON often describes records with child objects, such as customer.address.city, and lists of tags or contacts. Flattening turns that structure into columns named after their paths so the values can be used in CSV or a spreadsheet. This tool preserves scalar values and offers two ways to handle nested arrays. It does not validate a business schema, select a JSONPath expression or create a separate row for every item in a nested list.

Import a JSON file up to 30 MiB. The root must be an object or array; a standalone scalar such as 42 or null is rejected. A root object creates one row. A root array creates one row per element: scalar elements go into a value column, while objects are flattened. Everything runs in your browser. Native JSON parsing applies: comments and trailing commas are invalid, very large integers may lose precision, and repeated object keys retain their last value before flattening.

Nested object keys are joined with the chosen separator, a dot by default. An address object containing city becomes address.city. Existing key names do not escape that separator. If a literal address.city key collides with the equivalent nested path, processing fails rather than overwriting either value. Choose a separator absent from the original keys or rename them first. Empty objects are retained as the text {} rather than silently disappearing from the resulting table.

By default, a nested array stays in one cell as JSON text, for example ["blue","green"]. Index mode instead creates columns ending in 0, 1 and so forth, also traversing objects at those positions. It does not explode arrays into rows or form Cartesian products between lists. Empty arrays remain the text [] in either mode. Numbers, booleans and null retain scalar types; missing properties appear empty in tabular exports. Inspect the preview and export CSV or XLSX.