CSV to JSON Converter

The "trivially numeric" coercion rule, precisely: a field's trimmed value is converted to a JSON number only if it matches the pattern of a plain integer or decimal — optional leading minus sign, digits, optional decimal point and more digits — and does not have a leading zero (unless the value is exactly "0" or starts with "0."). This deliberately excludes values like "02139" (a US zip code) or "007" (which would otherwise become the misleading number 7), while still converting values like "30", "-12.5", or "0.75" to real numbers. Every other field, including anything with letters, symbols, or a leading-zero integer, stays a JSON string. This behaviour is a deliberate compromise, not full type inference — if you need guaranteed-string output for every field with no exceptions, quote the value in your CSV source and this tool leaves quoted-and-non-numeric-looking content as a string, which it already does by default. Empty fields (two consecutive commas, or a trailing comma before a line break) become an empty string `""` in the output, not `null` and not an omitted key — every row's object always has exactly the same set of keys as the header row, in the same order, which keeps the output predictable for downstream code that expects a fixed shape. Parsing details: quoted fields are recognised by a field starting with a double quote; everything up to the next unescaped double quote is taken literally, including commas and line breaks; `""` inside a quoted field becomes a single literal `"` in the output. Unquoted fields are simply the text between commas, trimmed of a leading/trailing carriage return but not otherwise trimmed of whitespace (so " London" with a leading space is preserved exactly, in case that space is meaningful in your data). Only the comma is treated as a field delimiter — semicolon-delimited or tab-delimited files are not auto-detected and should be converted to commas first, or handled with a different tool. The Download button saves the parsed result as a formatted (indented) `.json` file via a Blob and a temporary download link; the Copy button copies the same JSON text to your clipboard. If the header row contains duplicate column names, the later column silently overwrites the earlier one in each resulting object, since a JSON object cannot hold two values under the same key — rename duplicate columns in your CSV source if you need to keep both.