Updated 2026-10-05 · By Sergi (spnfury)
Turning a CSV into JSON sounds trivial until you hit a value with a comma inside quotes, a column of ZIP codes that must stay strings, or a 40 MB export that a site refuses to process. We ran the same three test files through six free converters and noted what each one got right.
The short version: CSVJSON and ConvertCSV are the most configurable, TableConvert is the best all-round editor, Online CSV Tools is convenient but gated, and CSV Creator (this site) is the simplest for a quick paste-and-copy. All six run in the browser without uploading your file, which should be the first thing you check.
Three inputs: a clean 5-row file; a file with quoted values containing commas, double quotes and accents ("López, Ana"); and a 25 MB export with 200,000 rows. For each tool we checked whether quoted fields survived, whether numbers and booleans were typed or left as strings, whether leading zeros were preserved, how nested keys (address.city) were handled, and whether the file was uploaded to a server (checked in the browser's network tab).
| Tool | Runs in browser | Quoted fields | Typed numbers | Nested keys | 25 MB file | Sign-up |
|---|---|---|---|---|---|---|
| CSV Creator | Yes | No (naive split) | No, all strings | No | Yes, a few seconds | No |
| CSVJSON (csvjson.com) | Yes | Yes | Optional | Yes, with dot notation | Yes | No |
| ConvertCSV | Yes | Yes | Optional | No | Slow but works | No |
| TableConvert | Yes | Yes | Optional | No | Yes | No (free tier) |
| Online CSV Tools | Yes | Yes | Optional | No | Limited on free plan | Premium for big files |
| Mr. Data Converter | Yes | Yes | Yes | No | Struggles | No |
Paste the CSV, click Convert, copy the pretty-printed array of objects. Keys come from the header row and every value is a string, so 08001 stays "08001" and nothing is reinterpreted. The limitation is the parser: it splits on commas and ignores quotes, so a value such as "Madrid, Spain" will spill into the next column. For clean exports it is the quickest option on this list; for messy files use one of the quote-aware tools below or clean the data first in the CSV Viewer and Editor.
Everything runs client-side, there are no ads inside the converter area, no sign-up and no size limit beyond your browser's memory. Output is JSON only; the reverse direction is a separate tool, JSON to CSV.
CSVJSON (csvjson.com) is open source and built by developers for developers. It parses quotes correctly, lets you choose between an array of objects or a hash keyed by the first column, can parse numbers automatically, and supports dot notation in headers to produce nested objects (address.city becomes {"address": {"city": ...}}). It also handles tab-separated input and lets you minify or pretty-print.
The interface is dense and the ads are visible, but the options are exactly the ones you need when the JSON has to match a schema.
ConvertCSV.com has been around for over a decade and converts CSV to JSON, SQL, XML, YAML and more. The CSV to JSON page offers input options (delimiter, quote character, first row as header), output options (array of objects, array of arrays, key/value) and templates for custom output. It handles quoted fields and large files, though the 25 MB test took noticeably longer than the others.
The design is dated and the page is heavy with ads, but if you need a specific output shape, ConvertCSV probably has a switch for it.
TableConvert.com shows the CSV in an editable grid before converting, which is handy for fixing a cell or deleting a column on the way to JSON. It outputs JSON in several shapes, as well as Markdown, LaTeX, SQL and HTML tables. Quoting is handled correctly and numbers can be typed.
The free tier covers normal use; very large files and some export formats sit behind a paid plan.
Online CSV Tools (by Browserling) is a collection of tiny single-purpose utilities with a clean interface. The CSV to JSON tool works well for small files, but the free plan limits input size and the site pushes a premium subscription.
Mr. Data Converter is a classic single-page tool from 2010 that still works: paste CSV, pick JSON (rows, columns or properties), copy. It types numbers automatically and handles quotes, but it has no options for large files and the page is no longer maintained.
All six tools in this comparison parse the file with JavaScript in your browser; we confirmed it in the network tab. That is not true of every converter online, so if privacy matters, open the developer tools, convert a small file and check that no request carries your data.
CSV has no types, so some converters keep every value as a string to avoid damaging codes with leading zeros. CSVJSON, ConvertCSV and TableConvert have an option to parse numbers; CSV Creator always keeps strings. Convert types in your code if you need numbers.
Use CSVJSON with headers in dot notation (customer.name, customer.email) and enable the nested output option. The other tools produce flat objects, one key per column.
An array of objects uses the header names as keys in every row ({"id": 1, "name": "Ana"}), which is what most APIs and JavaScript code expect. An array of arrays keeps only the values ([1, "Ana"]) and is smaller, but you need the header separately to know what each position means.
Yes. CSVJSON, ConvertCSV and TableConvert offer the reverse conversion on the same site, and CSV Creator has a separate JSON to CSV tool. Nested JSON must be flattened first in all of them except CSVJSON.