JSON Null Remover

About this tool

JSON Null Remover parses your JSON, walks the whole tree, and drops every object key whose value is null — at any depth, including objects nested inside arrays. What's left is re-serialized as valid JSON, pretty-printed or minified. Payloads from ORMs, form submissions and API responses shrink noticeably once the placeholder nulls are gone.

Emptiness beyond null is opt-in, one checkbox each: empty strings (""), empty arrays ([]) and empty objects ({}). Nothing else is touched — false and 0 are real values and are always kept, and the original key order is preserved.

Removal is bottom-up, so it cascades. Children are cleaned before their parent decides what to drop: with Also remove empty objects on, {"a": {"b": {"c": null}}} collapses all the way to {}, because each level is emptied in turn. With that box off, you get {"a": {"b": {}}} — the null is gone but the structure stays.

Worked example

Input:

{ "name": "Ada", "nickname": null, "address": { "city": "London", "zip": null } }

With the defaults (nulls only, Indent = 2), the output is:

{
  "name": "Ada",
  "address": {
    "city": "London"
  }
}

Now the same input with Also remove empty objects on and every value under address null — {"name": "Ada", "address": {"zip": null}} — collapses to {"name": "Ada"}: the emptied address object goes too.

Options

• Also remove empty strings — drops "" values. Off by default.
• Also remove empty arrays — drops [] values, including arrays emptied by the prune itself.
• Also remove empty objects — drops {} values, including objects emptied by the prune itself (this is what makes removal cascade upward).
• Trim whitespace from string values — trims each string first, so " " becomes "" and is removable when the empty-strings box is also ticked. On its own it just tidies the values.
• Values inside arrays — Compact (default) drops removable elements and closes the gap, so [1, null, 2] becomes [1, 2]. Keep leaves array elements exactly where they are, so indices stay stable for positional/tuple-shaped data; objects inside the array are still pruned.
• Indent spaces — 1–8 spaces per level (default 2), or 0 to minify onto one line.

Limits and behavior

• The input must be valid JSON. It's parsed before anything is removed, so a broken document is rejected with the parser's exact line and column — this tool prunes JSON, it doesn't repair it. Trailing commas, comments and unquoted keys are not accepted.
• The root value is never dropped. A document that prunes down to nothing comes back as {} (or []), never as an empty output, so the result is always valid JSON you can paste onwards.
• false, 0, "0" and " " (without trimming) are values, not emptiness — they always survive.
• Duplicate keys in the input resolve to the last occurrence, as in any strict JSON parser.
• Everything runs locally in WebAssembly, so document size is bounded only by your browser's memory; multi-megabyte documents work but take a moment to render.

FAQ