CSV to JSON Converter

STEP 01

Start with the receiving system, not the file extension

A spreadsheet is convenient for a human to scan, while an array of JSON objects is convenient for a program to consume. Conversion is useful when a content import, test fixture, or browser prototype expects named properties instead of columns. It is not a substitute for the destination's schema. Ask whether the importer expects an array, a single object, or a wrapper such as a records property before pasting the result.

This converter emits an array. It requires a header row and exactly the same column count in every record. It does not rename headers, trim field values, generate IDs, or infer nested objects. A column named customer.email remains a literal key, not an email property inside a customer object. These deliberate boundaries make a small import easier to audit.

STEP 02

Walk through an inventory example

Suppose an inventory export contains a padded SKU, a product label containing a comma, and a stock count. Choose comma as the separator and leave type conversion off for the first run. The quoted label remains one field and the SKU remains 001. Compare the number of objects with the data-row count, excluding the header, before saving the result.

If an API needs stock to be a number, enable safe scalar conversion and inspect the result again. In this tool only safe whole numbers and the exact lowercase literals true, false, and null are converted. Decimals, surrounding whitespace, padded digits, and integers outside JavaScript's safe range remain strings. That is a conservative tool choice, not a rule imposed by CSV or JSON.

sku,label,stock
001,"Notebook, blue",12

[
  {
    "sku": "001",
    "label": "Notebook, blue",
    "stock": "12"
  }
]

STEP 03

Troubleshoot the first rejected record

A column-count error usually means the selected delimiter is wrong or a value contains an unquoted separator. Do not fix it by deleting an arbitrary cell. Compare the rejected record with the header and inspect any commas inside labels. If the export uses semicolons or you pasted a tab-separated spreadsheet selection, change the delimiter and run again.

A literal quote inside a quoted field must be doubled. A line break can belong to a quoted field, so a record number is not always the same as a physical line number. This parser supports LF, CRLF, and CR line endings and removes one leading byte-order mark. Blank records are not silently discarded: fix them in the source when they do not match the column shape. Duplicate and blank headers are rejected rather than allowing values to disappear.

STEP 04

Validate the result before importing

After conversion, verify one ordinary record, one empty field, one multilingual label, and one unusually long identifier. Check that dates remain in the representation your destination expects. An empty cell becomes an empty string by default; it is not automatically a missing property or null. Download creates a UTF-8 JSON file in your browser, with no conversion upload.

Syntax correctness does not establish business correctness. A valid object can still contain an unknown field or an invalid date. Use the JSON Validator to inspect syntax, then test a small batch against your real import schema. For files above the local limit, use a streaming script or the destination's import service. Do not paste confidential records unless your organization's policy permits this browser-based workflow.

COMMON QUESTIONS

Before you use the result